prepare for merge

* Removed the sorting logic that forced column type to be sorted like
ColumnTypes.

* add comments

Co-authored-by: PSeitz <PSeitz@users.noreply.github.com>

Cargo.toml

@@ -55,11 +55,11 @@ measure_time = "0.8.2"
 async-trait = "0.1.53"
 arc-swap = "1.5.0"
 
+columnar = { version="0.1", path="./columnar", package ="tantivy-columnar" }
 sstable = { version="0.1", path="./sstable", package ="tantivy-sstable", optional = true }
 stacker = { version="0.1", path="./stacker", package ="tantivy-stacker" }
 tantivy-query-grammar = { version= "0.19.0", path="./query-grammar" }
 tantivy-bitpacker = 		{ version= "0.3", path="./bitpacker" }
-columnar = 		{ version= "0.1", path="./columnar", package="tantivy-columnar" }
 common = 								{ version= "0.5", path = "./common/", package = "tantivy-common" }
 fastfield_codecs = 			{ version= "0.3", path="./fastfield_codecs", default-features = false }
 tokenizer-api = { version="0.1", path="./tokenizer-api", package="tantivy-tokenizer-api" }

README.md

@@ -41,7 +41,7 @@ Your mileage WILL vary depending on the nature of queries and their load.
 - SIMD integer compression when the platform/CPU includes the SSE2 instruction set
 - Single valued and multivalued u64, i64, and f64 fast fields (equivalent of doc values in Lucene)
 - `&[u8]` fast fields
-- Text, i64, u64, f64, dates, and hierarchical facet fields
+- Text, i64, u64, f64, dates, ip, bool, and hierarchical facet fields
 - Compressed document store (LZ4, Zstd, None, Brotli, Snap)
 - Range queries
 - Faceted search
@@ -80,56 +80,21 @@ There are many ways to support this project.
 # Contributing code
 
 We use the GitHub Pull Request workflow: reference a GitHub ticket and/or include a comprehensive commit message when opening a PR.
+Feel free to update CHANGELOG.md with your contribution.
 
 ## Tokenizer
 
 When implementing a tokenizer for tantivy depend on the `tantivy-tokenizer-api` crate.
 
-## Minimum supported Rust version
-
-Tantivy currently requires at least Rust 1.62 or later to compile.
-
 ## Clone and build locally
 
 Tantivy compiles on stable Rust.
 To check out and run tests, you can simply run:
 
 ```bash
-    git clone https://github.com/quickwit-oss/tantivy.git
-    cd tantivy
-    cargo build
-```
-
-## Run tests
-
-Some tests will not run with just `cargo test` because of `fail-rs`.
-To run the tests exhaustively, run `./run-tests.sh`.
-
-## Debug
-
-You might find it useful to step through the programme with a debugger.
-
-### A failing test
-
-Make sure you haven't run `cargo clean` after the most recent `cargo test` or `cargo build` to guarantee that the `target/` directory exists. Use this bash script to find the name of the most recent debug build of Tantivy and run it under `rust-gdb`:
-
-```bash
-find target/debug/ -maxdepth 1 -executable -type f -name "tantivy*" -printf '%TY-%Tm-%Td %TT %p\n' | sort -r | cut -d " " -f 3 | xargs -I RECENT_DBG_TANTIVY rust-gdb RECENT_DBG_TANTIVY
-```
-
-Now that you are in `rust-gdb`, you can set breakpoints on lines and methods that match your source code and run the debug executable with flags that you normally pass to `cargo test` like this:
-
-```bash
-$gdb run --test-threads 1 --test $NAME_OF_TEST
-```
-
-### An example
-
-By default, `rustc` compiles everything in the `examples/` directory in debug mode. This makes it easy for you to make examples to reproduce bugs:
-
-```bash
-rust-gdb target/debug/examples/$EXAMPLE_NAME
-$ gdb run
+git clone https://github.com/quickwit-oss/tantivy.git
+cd tantivy
+cargo test
 ```
 
 # Companies Using Tantivy

TODO.txt

@@ -1,18 +0,0 @@
-Make schema_builder API fluent.
-fix doc serialization and prevent compression problems
-
-u64 , etc. shoudl return Resutl<Option> now that we support optional missing a column is really not an error
-remove fastfield codecs
-ditch the first_or_default trick. if it is still useful, improve its implementation.
-rename FastFieldReaders::open to load
-
-
-remove fast field reader
-
-find a way to unify the two DateTime.
-readd type check in the filter wrapper
-
-add unit test on columnar list columns.
-
-make sure sort works
-

bitpacker/Cargo.toml

@@ -15,3 +15,7 @@ homepage = "https://github.com/quickwit-oss/tantivy"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
+
+[dev-dependencies]
+rand = "0.8"
+proptest = "1"

bitpacker/benches/bench.rs

@@ -4,9 +4,39 @@ extern crate test;
 
 #[cfg(test)]
 mod tests {
-    use tantivy_bitpacker::BlockedBitpacker;
+    use rand::seq::IteratorRandom;
+    use rand::thread_rng;
+    use tantivy_bitpacker::{BitPacker, BitUnpacker, BlockedBitpacker};
     use test::Bencher;
 
+    #[inline(never)]
+    fn create_bitpacked_data(bit_width: u8, num_els: u32) -> Vec<u8> {
+        let mut bitpacker = BitPacker::new();
+        let mut buffer = Vec::new();
+        for _ in 0..num_els {
+            // the values do not matter.
+            bitpacker.write(0u64, bit_width, &mut buffer).unwrap();
+            bitpacker.flush(&mut buffer).unwrap();
+        }
+        buffer
+    }
+
+    #[bench]
+    fn bench_bitpacking_read(b: &mut Bencher) {
+        let bit_width = 3;
+        let num_els = 1_000_000u32;
+        let bit_unpacker = BitUnpacker::new(bit_width);
+        let data = create_bitpacked_data(bit_width, num_els);
+        let idxs: Vec<u32> = (0..num_els).choose_multiple(&mut thread_rng(), 100_000);
+        b.iter(|| {
+            let mut out = 0u64;
+            for &idx in &idxs {
+                out = out.wrapping_add(bit_unpacker.get(idx, &data[..]));
+            }
+            out
+        });
+    }
+
     #[bench]
     fn bench_blockedbitp_read(b: &mut Bencher) {
         let mut blocked_bitpacker = BlockedBitpacker::new();
@@ -14,9 +44,9 @@ mod tests {
             blocked_bitpacker.add(val * val);
         }
         b.iter(|| {
-            let mut out = 0;
+            let mut out = 0u64;
             for val in 0..=21500 {
-                out = blocked_bitpacker.get(val);
+                out = out.wrapping_add(blocked_bitpacker.get(val));
             }
             out
         });

bitpacker/src/bitpacker.rs

@@ -56,27 +56,31 @@ impl BitPacker {
 
     pub fn close<TWrite: io::Write>(&mut self, output: &mut TWrite) -> io::Result<()> {
         self.flush(output)?;
-        // Padding the write file to simplify reads.
-        output.write_all(&[0u8; 7])?;
         Ok(())
     }
 }
 
-#[derive(Clone, Debug, Default)]
+#[derive(Clone, Debug, Default, Copy)]
 pub struct BitUnpacker {
-    num_bits: u64,
+    num_bits: u32,
     mask: u64,
 }
 
 impl BitUnpacker {
+    /// Creates a bit unpacker, that assumes the same bitwidth for all values.
+    ///
+    /// The bitunpacker works by doing an unaligned read of 8 bytes.
+    /// For this reason, values of `num_bits` between
+    /// [57..63] are forbidden.
     pub fn new(num_bits: u8) -> BitUnpacker {
+        assert!(num_bits <= 7 * 8 || num_bits == 64);
         let mask: u64 = if num_bits == 64 {
             !0u64
         } else {
             (1u64 << num_bits) - 1u64
         };
         BitUnpacker {
-            num_bits: u64::from(num_bits),
+            num_bits: u32::from(num_bits),
             mask,
         }
     }
@@ -87,28 +91,40 @@ impl BitUnpacker {
 
     #[inline]
     pub fn get(&self, idx: u32, data: &[u8]) -> u64 {
-        if self.num_bits == 0 {
-            return 0u64;
-        }
-        let addr_in_bits = idx * self.num_bits as u32;
+        let addr_in_bits = idx * self.num_bits;
         let addr = (addr_in_bits >> 3) as usize;
+        if addr + 8 > data.len() {
+            if self.num_bits == 0 {
+                return 0;
+            }
+            let bit_shift = addr_in_bits & 7;
+            return self.get_slow_path(addr, bit_shift, data);
+        }
         let bit_shift = addr_in_bits & 7;
-        debug_assert!(
-            addr + 8 <= data.len(),
-            "The fast field field should have been padded with 7 bytes."
-        );
         let bytes: [u8; 8] = (&data[addr..addr + 8]).try_into().unwrap();
         let val_unshifted_unmasked: u64 = u64::from_le_bytes(bytes);
         let val_shifted = val_unshifted_unmasked >> bit_shift;
         val_shifted & self.mask
     }
+
+    #[inline(never)]
+    fn get_slow_path(&self, addr: usize, bit_shift: u32, data: &[u8]) -> u64 {
+        let mut bytes: [u8; 8] = [0u8; 8];
+        let available_bytes = data.len() - addr;
+        // This function is meant to only be called if we did not have 8 bytes to load.
+        debug_assert!(available_bytes < 8);
+        bytes[..available_bytes].copy_from_slice(&data[addr..]);
+        let val_unshifted_unmasked: u64 = u64::from_le_bytes(bytes);
+        let val_shifted = val_unshifted_unmasked >> bit_shift;
+        val_shifted & self.mask
+    }
 }
 
 #[cfg(test)]
 mod test {
     use super::{BitPacker, BitUnpacker};
 
-    fn create_fastfield_bitpacker(len: usize, num_bits: u8) -> (BitUnpacker, Vec<u64>, Vec<u8>) {
+    fn create_bitpacker(len: usize, num_bits: u8) -> (BitUnpacker, Vec<u64>, Vec<u8>) {
         let mut data = Vec::new();
         let mut bitpacker = BitPacker::new();
         let max_val: u64 = (1u64 << num_bits as u64) - 1u64;
@@ -119,13 +135,13 @@ mod test {
             bitpacker.write(val, num_bits, &mut data).unwrap();
         }
         bitpacker.close(&mut data).unwrap();
-        assert_eq!(data.len(), ((num_bits as usize) * len + 7) / 8 + 7);
+        assert_eq!(data.len(), ((num_bits as usize) * len + 7) / 8);
         let bitunpacker = BitUnpacker::new(num_bits);
         (bitunpacker, vals, data)
     }
 
     fn test_bitpacker_util(len: usize, num_bits: u8) {
-        let (bitunpacker, vals, data) = create_fastfield_bitpacker(len, num_bits);
+        let (bitunpacker, vals, data) = create_bitpacker(len, num_bits);
         for (i, val) in vals.iter().enumerate() {
             assert_eq!(bitunpacker.get(i as u32, &data), *val);
         }
@@ -139,4 +155,49 @@ mod test {
         test_bitpacker_util(6, 14);
         test_bitpacker_util(1000, 14);
     }
+
+    use proptest::prelude::*;
+
+    fn num_bits_strategy() -> impl Strategy<Value = u8> {
+        prop_oneof!(Just(0), Just(1), 2u8..56u8, Just(56), Just(64),)
+    }
+
+    fn vals_strategy() -> impl Strategy<Value = (u8, Vec<u64>)> {
+        (num_bits_strategy(), 0usize..100usize).prop_flat_map(|(num_bits, len)| {
+            let max_val = if num_bits == 64 {
+                u64::MAX
+            } else {
+                (1u64 << num_bits as u32) - 1
+            };
+            let vals = proptest::collection::vec(0..=max_val, len);
+            vals.prop_map(move |vals| (num_bits, vals))
+        })
+    }
+
+    fn test_bitpacker_aux(num_bits: u8, vals: &[u64]) {
+        let mut buffer: Vec<u8> = Vec::new();
+        let mut bitpacker = BitPacker::new();
+        for &val in vals {
+            bitpacker.write(val, num_bits, &mut buffer).unwrap();
+        }
+        bitpacker.flush(&mut buffer).unwrap();
+        assert_eq!(buffer.len(), (vals.len() * num_bits as usize + 7) / 8);
+        let bitunpacker = BitUnpacker::new(num_bits);
+        let max_val = if num_bits == 64 {
+            u64::MAX
+        } else {
+            (1u64 << num_bits) - 1
+        };
+        for (i, val) in vals.iter().copied().enumerate() {
+            assert!(val <= max_val);
+            assert_eq!(bitunpacker.get(i as u32, &buffer), val);
+        }
+    }
+
+    proptest::proptest! {
+        #[test]
+        fn test_bitpacker_proptest((num_bits, vals) in vals_strategy()) {
+            test_bitpacker_aux(num_bits, &vals);
+        }
+    }
 }

columnar/Cargo.toml

@@ -5,24 +5,23 @@ edition = "2021"
 license = "MIT"
 
 [dependencies]
+itertools = "0.10.5"
+log = "0.4.17"
+fnv = "1.0.7"
+fastdivide = "0.4.0"
+rand = { version = "0.8.5", optional = true }
+measure_time = { version = "0.8.2", optional = true }
+prettytable-rs = { version = "0.10.0", optional = true }
+
 stacker = { path = "../stacker", package="tantivy-stacker"}
-serde_json = "1"
-thiserror = "1"
-fnv = "1"
 sstable = { path = "../sstable", package = "tantivy-sstable" }
 common = { path = "../common", package = "tantivy-common" }
-itertools = "0.10"
-log = "0.4"
 tantivy-bitpacker = { version= "0.3", path = "../bitpacker/" }
-prettytable-rs = {version="0.10.0", optional= true}
-rand = {version="0.8.3", optional= true}
-fastdivide = "0.4"
-measure_time = { version="0.8.2", optional=true}
 
 [dev-dependencies]
-proptest = "1"
-more-asserts = "0.3.0"
-rand = "0.8.3"
+proptest = "1.0.0"
+more-asserts = "0.3.1"
+rand = "0.8.5"
 
 [features]
 unstable = []

columnar/src/TODO.md

@@ -26,7 +26,6 @@ Add alignment?
 Consider another codec to bridge the gap between few and 5k elements
 
 # Cleanup and rationalization
-remove the 6 bit limitation of columntype. use 4 + 4 bits instead.
 in benchmark, unify percent vs ratio, f32 vs f64.
 investigate if should have better errors? io::Error is overused at the moment.
 rename rank/select in unit tests

columnar/src/column/dictionary_encoded.rs

@@ -5,10 +5,16 @@ use std::sync::Arc;
 use sstable::{Dictionary, VoidSSTable};
 
 use crate::column::Column;
-use crate::column_index::ColumnIndex;
 use crate::RowId;
 
 /// Dictionary encoded column.
+///
+/// The column simply gives access to a regular u64-column that, in
+/// which the values are term-ordinals.
+///
+/// These ordinals are ids uniquely identify the bytes that are stored in
+/// the column. These ordinals are small, and sorted in the same order
+/// as the term_ord_column.
 #[derive(Clone)]
 pub struct BytesColumn {
     pub(crate) dictionary: Arc<Dictionary<VoidSSTable>>,
@@ -16,30 +22,57 @@ pub struct BytesColumn {
 }
 
 impl BytesColumn {
+    /// Fills the given `output` buffer with the term associated to the ordinal `ord`.
+    ///
     /// Returns `false` if the term does not exist (e.g. `term_ord` is greater or equal to the
     /// overll number of terms).
-    pub fn term_ord_to_str(&self, term_ord: u64, output: &mut Vec<u8>) -> io::Result<bool> {
-        self.dictionary.ord_to_term(term_ord, output)
-    }
-
-    pub fn term_ords(&self) -> &Column<u64> {
-        &self.term_ord_column
+    pub fn ord_to_bytes(&self, ord: u64, output: &mut Vec<u8>) -> io::Result<bool> {
+        self.dictionary.ord_to_term(ord, output)
     }
 
+    /// Returns the number of rows in the column.
     pub fn num_rows(&self) -> RowId {
         self.term_ord_column.num_rows()
     }
-}
 
-impl Deref for BytesColumn {
-    type Target = ColumnIndex<'static>;
-
-    fn deref(&self) -> &Self::Target {
-        &**self.term_ords()
+    /// Returns the column of ordinals
+    pub fn ords(&self) -> &Column<u64> {
+        &self.term_ord_column
     }
 }
 
-#[cfg(test)]
-mod tests {
-    use crate::{ColumnarReader, ColumnarWriter};
+#[derive(Clone)]
+pub struct StrColumn(BytesColumn);
+
+impl From<BytesColumn> for StrColumn {
+    fn from(bytes_col: BytesColumn) -> Self {
+        StrColumn(bytes_col)
+    }
+}
+
+impl StrColumn {
+    /// Fills the buffer
+    pub fn ord_to_str(&self, term_ord: u64, output: &mut String) -> io::Result<bool> {
+        unsafe {
+            let buf = output.as_mut_vec();
+            self.0.dictionary.ord_to_term(term_ord, buf)?;
+            // TODO consider remove checks if it hurts performance.
+            if std::str::from_utf8(buf.as_slice()).is_err() {
+                buf.clear();
+                return Err(io::Error::new(
+                    io::ErrorKind::InvalidData,
+                    "Not valid utf-8",
+                ));
+            }
+        }
+        Ok(true)
+    }
+}
+
+impl Deref for StrColumn {
+    type Target = BytesColumn;
+
+    fn deref(&self) -> &Self::Target {
+        &self.0
+    }
 }

columnar/src/column/mod.rs

@@ -5,10 +5,13 @@ use std::ops::Deref;
 use std::sync::Arc;
 
 use common::BinarySerializable;
-pub use dictionary_encoded::BytesColumn;
-pub use serialize::{open_column_bytes, open_column_u64, serialize_column_u64};
+pub use dictionary_encoded::{BytesColumn, StrColumn};
+pub use serialize::{
+    open_column_bytes, open_column_u128, open_column_u64, serialize_column_mappable_to_u128,
+    serialize_column_mappable_to_u64,
+};
 
-use crate::column_index::{ColumnIndex, Set};
+use crate::column_index::ColumnIndex;
 use crate::column_values::ColumnValues;
 use crate::{Cardinality, RowId};
 
@@ -19,27 +22,37 @@ pub struct Column<T> {
 }
 
 impl<T: PartialOrd> Column<T> {
+    pub fn get_cardinality(&self) -> Cardinality {
+        self.idx.get_cardinality()
+    }
     pub fn num_rows(&self) -> RowId {
         match &self.idx {
-            ColumnIndex::Full => self.values.num_vals(),
-            ColumnIndex::Optional(optional_idx) => optional_idx.num_rows(),
-            ColumnIndex::Multivalued(_) => todo!(),
+            ColumnIndex::Full => self.values.num_vals() as u32,
+            ColumnIndex::Optional(optional_index) => optional_index.num_rows(),
+            ColumnIndex::Multivalued(col_index) => {
+                // The multivalued index contains all value start row_id,
+                // and one extra value at the end with the overall number of rows.
+                col_index.num_vals() - 1
+            }
         }
     }
+
+    pub fn min_value(&self) -> T {
+        self.values.min_value()
+    }
+    pub fn max_value(&self) -> T {
+        self.values.max_value()
+    }
 }
 
 impl<T: PartialOrd + Copy + Send + Sync + 'static> Column<T> {
     pub fn first(&self, row_id: RowId) -> Option<T> {
-        match &self.idx {
-            ColumnIndex::Full => Some(self.values.get_val(row_id)),
-            ColumnIndex::Optional(opt_idx) => {
-                let value_row_idx = opt_idx.rank_if_exists(row_id)?;
-                Some(self.values.get_val(value_row_idx))
-            }
-            ColumnIndex::Multivalued(_multivalued_index) => {
-                todo!();
-            }
-        }
+        self.values(row_id).next()
+    }
+
+    pub fn values(&self, row_id: RowId) -> impl Iterator<Item = T> + '_ {
+        self.value_row_ids(row_id)
+            .map(|value_row_id: RowId| self.values.get_val(value_row_id))
     }
 
     pub fn first_or_default_col(self, default_value: T) -> Arc<dyn ColumnValues<T>> {

columnar/src/column/serialize.rs

@@ -2,24 +2,51 @@ use std::io;
 use std::io::Write;
 use std::sync::Arc;
 
-use common::{CountingWriter, OwnedBytes};
+use common::OwnedBytes;
 use sstable::Dictionary;
 
 use crate::column::{BytesColumn, Column};
 use crate::column_index::{serialize_column_index, SerializableColumnIndex};
+use crate::column_values::serialize::serialize_column_values_u128;
 use crate::column_values::{
-    serialize_column_values, ColumnValues, MonotonicallyMappableToU64, ALL_CODEC_TYPES,
+    serialize_column_values, ColumnValues, FastFieldCodecType, MonotonicallyMappableToU128,
+    MonotonicallyMappableToU64,
 };
-pub fn serialize_column_u64<T: MonotonicallyMappableToU64>(
+
+pub fn serialize_column_mappable_to_u128<
+    F: Fn() -> I,
+    I: Iterator<Item = T>,
+    T: MonotonicallyMappableToU128,
+>(
+    column_index: SerializableColumnIndex<'_>,
+    column_values: F,
+    num_vals: u32,
+    output: &mut impl Write,
+) -> io::Result<()> {
+    let column_index_num_bytes = serialize_column_index(column_index, output)?;
+    serialize_column_values_u128(
+        || column_values().map(|val| val.to_u128()),
+        num_vals,
+        output,
+    )?;
+    output.write_all(&column_index_num_bytes.to_le_bytes())?;
+    Ok(())
+}
+
+pub fn serialize_column_mappable_to_u64<T: MonotonicallyMappableToU64>(
     column_index: SerializableColumnIndex<'_>,
     column_values: &impl ColumnValues<T>,
     output: &mut impl Write,
 ) -> io::Result<()> {
-    let mut counting_writer = CountingWriter::wrap(output);
-    serialize_column_index(column_index, &mut counting_writer)?;
-    let column_index_num_bytes = counting_writer.written_bytes() as u32;
-    let output = counting_writer.finish();
-    serialize_column_values(column_values, &ALL_CODEC_TYPES[..], output)?;
+    let column_index_num_bytes = serialize_column_index(column_index, output)?;
+    serialize_column_values(
+        column_values,
+        &[
+            FastFieldCodecType::Bitpacked,
+            FastFieldCodecType::BlockwiseLinear,
+        ],
+        output,
+    )?;
     output.write_all(&column_index_num_bytes.to_le_bytes())?;
     Ok(())
 }
@@ -41,14 +68,34 @@ pub fn open_column_u64<T: MonotonicallyMappableToU64>(bytes: OwnedBytes) -> io::
     })
 }
 
-pub fn open_column_bytes(data: OwnedBytes) -> io::Result<BytesColumn> {
+pub fn open_column_u128<T: MonotonicallyMappableToU128>(
+    bytes: OwnedBytes,
+) -> io::Result<Column<T>> {
+    let (body, column_index_num_bytes_payload) = bytes.rsplit(4);
+    let column_index_num_bytes = u32::from_le_bytes(
+        column_index_num_bytes_payload
+            .as_slice()
+            .try_into()
+            .unwrap(),
+    );
+    let (column_index_data, column_values_data) = body.split(column_index_num_bytes as usize);
+    let column_index = crate::column_index::open_column_index(column_index_data)?;
+    let column_values = crate::column_values::open_u128_mapped(column_values_data)?;
+    Ok(Column {
+        idx: column_index,
+        values: column_values,
+    })
+}
+
+pub fn open_column_bytes<T: From<BytesColumn>>(data: OwnedBytes) -> io::Result<T> {
     let (body, dictionary_len_bytes) = data.rsplit(4);
     let dictionary_len = u32::from_le_bytes(dictionary_len_bytes.as_slice().try_into().unwrap());
     let (dictionary_bytes, column_bytes) = body.split(dictionary_len as usize);
     let dictionary = Arc::new(Dictionary::from_bytes(dictionary_bytes)?);
     let term_ord_column = crate::column::open_column_u64::<u64>(column_bytes)?;
-    Ok(BytesColumn {
+    let bytes_column = BytesColumn {
         dictionary,
         term_ord_column,
-    })
+    };
+    Ok(bytes_column.into())
 }

columnar/src/column_index/mod.rs

@@ -2,6 +2,7 @@ mod multivalued_index;
 mod optional_index;
 mod serialize;
 
+use std::ops::Range;
 use std::sync::Arc;
 
 pub use optional_index::{OptionalIndex, SerializableOptionalIndex, Set};
@@ -14,8 +15,12 @@ use crate::{Cardinality, RowId};
 pub enum ColumnIndex<'a> {
     Full,
     Optional(OptionalIndex),
-    // TODO remove the Arc<dyn> apart from serialization this is not
-    // dynamic at all.
+    // TODO Remove the static by fixing the codec if possible.
+    /// The column values enclosed contains for all row_id,
+    /// the value start_index.
+    ///
+    /// In addition, at index num_rows, an extra value is added
+    /// containing the overal number of values.
     Multivalued(Arc<dyn ColumnValues<RowId> + 'a>),
 }
 
@@ -27,4 +32,23 @@ impl<'a> ColumnIndex<'a> {
             ColumnIndex::Multivalued(_) => Cardinality::Multivalued,
         }
     }
+
+    pub fn value_row_ids(&self, row_id: RowId) -> Range<RowId> {
+        match self {
+            ColumnIndex::Full => row_id..row_id + 1,
+            ColumnIndex::Optional(optional_index) => {
+                if let Some(val) = optional_index.rank_if_exists(row_id) {
+                    val..val + 1
+                } else {
+                    0..0
+                }
+            }
+            ColumnIndex::Multivalued(multivalued_index) => {
+                let multivalued_index_ref = &**multivalued_index;
+                let start: u32 = multivalued_index_ref.get_val(row_id);
+                let end: u32 = multivalued_index_ref.get_val(row_id + 1);
+                start..end
+            }
+        }
+    }
 }

columnar/src/column_index/multivalued_index.rs

@@ -11,11 +11,11 @@ use crate::RowId;
 pub struct MultivaluedIndex(Arc<dyn ColumnValues<RowId>>);
 
 pub fn serialize_multivalued_index(
-    multivalued_index: MultivaluedIndex,
+    multivalued_index: &dyn ColumnValues<RowId>,
     output: &mut impl Write,
 ) -> io::Result<()> {
     crate::column_values::serialize_column_values(
-        &*multivalued_index.0,
+        &*multivalued_index,
         &[FastFieldCodecType::Bitpacked, FastFieldCodecType::Linear],
         output,
     )?;
@@ -23,5 +23,7 @@ pub fn serialize_multivalued_index(
 }
 
 pub fn open_multivalued_index(bytes: OwnedBytes) -> io::Result<Arc<dyn ColumnValues<RowId>>> {
-    todo!();
+    let start_index_column: Arc<dyn ColumnValues<RowId>> =
+        crate::column_values::open_u64_mapped(bytes)?;
+    Ok(start_index_column)
 }

columnar/src/column_index/optional_index/set_block/mod.rs

@@ -1,7 +1,7 @@
-mod set_block;
+mod dense;
 mod sparse;
 
-pub use set_block::{DenseBlock, DenseBlockCodec, DENSE_BLOCK_NUM_BYTES};
+pub use dense::{DenseBlock, DenseBlockCodec, DENSE_BLOCK_NUM_BYTES};
 pub use sparse::{SparseBlock, SparseBlockCodec};
 
 #[cfg(test)]

columnar/src/column_index/optional_index/set_block/tests.rs

@@ -1,7 +1,8 @@
 use std::collections::HashMap;
 
-use crate::column_index::optional_index::set_block::set_block::DENSE_BLOCK_NUM_BYTES;
-use crate::column_index::optional_index::set_block::{DenseBlockCodec, SparseBlockCodec};
+use crate::column_index::optional_index::set_block::{
+    DenseBlockCodec, SparseBlockCodec, DENSE_BLOCK_NUM_BYTES,
+};
 use crate::column_index::optional_index::{Set, SetCodec};
 
 fn test_set_helper<C: SetCodec<Item = u16>>(vals: &[u16]) -> usize {
@@ -51,6 +52,7 @@ fn test_sparse_block_set_u16_max() {
 use proptest::prelude::*;
 
 proptest! {
+    #![proptest_config(ProptestConfig::with_cases(1))]
     #[test]
     fn test_prop_test_dense(els in proptest::collection::btree_set(0..=u16::MAX, 0..=u16::MAX as usize)) {
         let vals: Vec<u16> = els.into_iter().collect();

columnar/src/column_index/serialize.rs

@@ -1,19 +1,20 @@
 use std::io;
 use std::io::Write;
 
-use common::OwnedBytes;
+use common::{CountingWriter, OwnedBytes};
 
-use crate::column_index::multivalued_index::{serialize_multivalued_index, MultivaluedIndex};
+use crate::column_index::multivalued_index::serialize_multivalued_index;
 use crate::column_index::optional_index::serialize_optional_index;
 use crate::column_index::{ColumnIndex, SerializableOptionalIndex};
-use crate::Cardinality;
+use crate::column_values::ColumnValues;
+use crate::{Cardinality, RowId};
 
 pub enum SerializableColumnIndex<'a> {
     Full,
     Optional(Box<dyn SerializableOptionalIndex<'a> + 'a>),
     // TODO remove the Arc<dyn> apart from serialization this is not
     // dynamic at all.
-    Multivalued(MultivaluedIndex),
+    Multivalued(Box<dyn ColumnValues<RowId> + 'a>),
 }
 
 impl<'a> SerializableColumnIndex<'a> {
@@ -29,19 +30,21 @@ impl<'a> SerializableColumnIndex<'a> {
 pub fn serialize_column_index(
     column_index: SerializableColumnIndex,
     output: &mut impl Write,
-) -> io::Result<()> {
+) -> io::Result<u32> {
+    let mut output = CountingWriter::wrap(output);
     let cardinality = column_index.get_cardinality().to_code();
     output.write_all(&[cardinality])?;
     match column_index {
         SerializableColumnIndex::Full => {}
         SerializableColumnIndex::Optional(optional_index) => {
-            serialize_optional_index(&*optional_index, output)?
+            serialize_optional_index(&*optional_index, &mut output)?
         }
         SerializableColumnIndex::Multivalued(multivalued_index) => {
-            serialize_multivalued_index(multivalued_index, output)?
+            serialize_multivalued_index(&*multivalued_index, &mut output)?
         }
     }
-    Ok(())
+    let column_index_num_bytes = output.written_bytes() as u32;
+    Ok(column_index_num_bytes)
 }
 
 pub fn open_column_index(mut bytes: OwnedBytes) -> io::Result<ColumnIndex<'static>> {

columnar/src/column_values/column.rs

@@ -78,6 +78,32 @@ pub trait ColumnValues<T: PartialOrd = u64>: Send + Sync {
     }
 }
 
+impl<T: Copy + PartialOrd> ColumnValues<T> for std::sync::Arc<dyn ColumnValues<T>> {
+    fn get_val(&self, idx: u32) -> T {
+        self.as_ref().get_val(idx)
+    }
+
+    fn min_value(&self) -> T {
+        self.as_ref().min_value()
+    }
+
+    fn max_value(&self) -> T {
+        self.as_ref().max_value()
+    }
+
+    fn num_vals(&self) -> u32 {
+        self.as_ref().num_vals()
+    }
+
+    fn iter<'b>(&'b self) -> Box<dyn Iterator<Item = T> + 'b> {
+        self.as_ref().iter()
+    }
+
+    fn get_range(&self, start: u64, output: &mut [T]) {
+        self.as_ref().get_range(start, output)
+    }
+}
+
 impl<'a, C: ColumnValues<T> + ?Sized, T: Copy + PartialOrd> ColumnValues<T> for &'a C {
     fn get_val(&self, idx: u32) -> T {
         (*self).get_val(idx)

columnar/src/column_values/mod.rs

@@ -28,7 +28,7 @@ mod compact_space;
 mod line;
 mod linear;
 pub(crate) mod monotonic_mapping;
-// mod monotonic_mapping_u128;
+pub(crate) mod monotonic_mapping_u128;
 
 mod column;
 mod column_with_cardinality;
@@ -37,8 +37,10 @@ pub mod serialize;
 
 pub use self::column::{monotonic_map_column, ColumnValues, IterColumn, VecColumn};
 pub use self::monotonic_mapping::{MonotonicallyMappableToU64, StrictlyMonotonicFn};
-// pub use self::monotonic_mapping_u128::MonotonicallyMappableToU128;
-pub use self::serialize::{serialize_and_load, serialize_column_values, NormalizedHeader};
+pub use self::monotonic_mapping_u128::MonotonicallyMappableToU128;
+#[cfg(test)]
+pub use self::serialize::tests::serialize_and_load;
+pub use self::serialize::{serialize_column_values, NormalizedHeader};
 use crate::column_values::bitpacked::BitpackedCodec;
 use crate::column_values::blockwise_linear::BlockwiseLinearCodec;
 use crate::column_values::linear::LinearCodec;
@@ -122,19 +124,17 @@ impl U128FastFieldCodecType {
 }
 
 /// Returns the correct codec reader wrapped in the `Arc` for the data.
-// pub fn open_u128<Item: MonotonicallyMappableToU128>(
-//     bytes: OwnedBytes,
-// ) -> io::Result<Arc<dyn Column<Item>>> {
-//     todo!();
-//     // let (bytes, _format_version) = read_format_version(bytes)?;
-//     // let (mut bytes, _null_index_footer) = read_null_index_footer(bytes)?;
-//     // let header = U128Header::deserialize(&mut bytes)?;
-//     // assert_eq!(header.codec_type, U128FastFieldCodecType::CompactSpace);
-//     // let reader = CompactSpaceDecompressor::open(bytes)?;
-//     // let inverted: StrictlyMonotonicMappingInverter<StrictlyMonotonicMappingToInternal<Item>> =
-//     //     StrictlyMonotonicMappingToInternal::<Item>::new().into();
-//     // Ok(Arc::new(monotonic_map_column(reader, inverted)))
-// }
+pub fn open_u128_mapped<T: MonotonicallyMappableToU128>(
+    mut bytes: OwnedBytes,
+) -> io::Result<Arc<dyn ColumnValues<T>>> {
+    let header = U128Header::deserialize(&mut bytes)?;
+    assert_eq!(header.codec_type, U128FastFieldCodecType::CompactSpace);
+    let reader = CompactSpaceDecompressor::open(bytes)?;
+
+    let inverted: StrictlyMonotonicMappingInverter<StrictlyMonotonicMappingToInternal<T>> =
+        StrictlyMonotonicMappingToInternal::<T>::new().into();
+    Ok(Arc::new(monotonic_map_column(reader, inverted)))
+}
 
 /// Returns the correct codec reader wrapped in the `Arc` for the data.
 pub fn open_u64_mapped<T: MonotonicallyMappableToU64>(
@@ -198,13 +198,6 @@ pub(crate) trait FastFieldCodec: 'static {
     fn estimate(column: &dyn ColumnValues) -> Option<f32>;
 }
 
-/// The list of all available codecs for u64 convertible data.
-pub const ALL_CODEC_TYPES: [FastFieldCodecType; 3] = [
-    FastFieldCodecType::Bitpacked,
-    FastFieldCodecType::BlockwiseLinear,
-    FastFieldCodecType::Linear,
-];
-
 #[cfg(all(test, feature = "unstable"))]
 mod bench {
     use std::sync::Arc;

columnar/src/column_values/monotonic_mapping.rs

@@ -2,6 +2,7 @@ use std::marker::PhantomData;
 
 use fastdivide::DividerU64;
 
+use super::MonotonicallyMappableToU128;
 use crate::RowId;
 
 /// Monotonic maps a value to u64 value space.
@@ -80,21 +81,20 @@ impl<T> StrictlyMonotonicMappingToInternal<T> {
     }
 }
 
-// TODO
-// impl<External: MonotonicallyMappableToU128, T: MonotonicallyMappableToU128>
-//     StrictlyMonotonicFn<External, u128> for StrictlyMonotonicMappingToInternal<T>
-// where T: MonotonicallyMappableToU128
-// {
-//     #[inline(always)]
-//     fn mapping(&self, inp: External) -> u128 {
-//         External::to_u128(inp)
-//     }
+impl<External: MonotonicallyMappableToU128, T: MonotonicallyMappableToU128>
+    StrictlyMonotonicFn<External, u128> for StrictlyMonotonicMappingToInternal<T>
+where T: MonotonicallyMappableToU128
+{
+    #[inline(always)]
+    fn mapping(&self, inp: External) -> u128 {
+        External::to_u128(inp)
+    }
 
-//     #[inline(always)]
-//     fn inverse(&self, out: u128) -> External {
-//         External::from_u128(out)
-//     }
-// }
+    #[inline(always)]
+    fn inverse(&self, out: u128) -> External {
+        External::from_u128(out)
+    }
+}
 
 impl<External: MonotonicallyMappableToU64, T: MonotonicallyMappableToU64>
     StrictlyMonotonicFn<External, u64> for StrictlyMonotonicMappingToInternal<T>

columnar/src/column_values/serialize.rs

@@ -19,9 +19,8 @@
 
 use std::io;
 use std::num::NonZeroU64;
-use std::sync::Arc;
 
-use common::{BinarySerializable, OwnedBytes, VInt};
+use common::{BinarySerializable, VInt};
 use log::warn;
 
 use super::bitpacked::BitpackedCodec;
@@ -33,8 +32,9 @@ use super::monotonic_mapping::{
 };
 use super::{
     monotonic_map_column, ColumnValues, FastFieldCodec, FastFieldCodecType,
-    MonotonicallyMappableToU64, U128FastFieldCodecType, VecColumn, ALL_CODEC_TYPES,
+    MonotonicallyMappableToU64, U128FastFieldCodecType,
 };
+use crate::column_values::compact_space::CompactSpaceCompressor;
 
 /// The normalized header gives some parameters after applying the following
 /// normalization of the vector:
@@ -160,54 +160,22 @@ impl BinarySerializable for Header {
     }
 }
 
-/// Return estimated compression for given codec in the value range [0.0..1.0], where 1.0 means no
-/// compression.
-pub(crate) fn estimate<T: MonotonicallyMappableToU64>(
-    typed_column: impl ColumnValues<T>,
-    codec_type: FastFieldCodecType,
-) -> Option<f32> {
-    let column = monotonic_map_column(typed_column, StrictlyMonotonicMappingToInternal::<T>::new());
-    let min_value = column.min_value();
-    let gcd = super::gcd::find_gcd(column.iter().map(|val| val - min_value))
-        .filter(|gcd| gcd.get() > 1u64);
-    let mapping = StrictlyMonotonicMappingToInternalGCDBaseval::new(
-        gcd.map(|gcd| gcd.get()).unwrap_or(1u64),
-        min_value,
-    );
-    let normalized_column = monotonic_map_column(&column, mapping);
-    match codec_type {
-        FastFieldCodecType::Bitpacked => BitpackedCodec::estimate(&normalized_column),
-        FastFieldCodecType::Linear => LinearCodec::estimate(&normalized_column),
-        FastFieldCodecType::BlockwiseLinear => BlockwiseLinearCodec::estimate(&normalized_column),
-    }
-}
-
-// TODO
 /// Serializes u128 values with the compact space codec.
-// pub fn serialize_u128_new<F: Fn() -> I, I: Iterator<Item = u128>>(
-//     value_index: ColumnIndex,
-//     iter_gen: F,
-//     num_vals: u32,
-//     output: &mut impl io::Write,
-// ) -> io::Result<()> {
-//     let header = U128Header {
-//         num_vals,
-//         codec_type: U128FastFieldCodecType::CompactSpace,
-//     };
-//     header.serialize(output)?;
-//     let compressor = CompactSpaceCompressor::train_from(iter_gen(), num_vals);
-//     compressor.compress_into(iter_gen(), output).unwrap();
+pub fn serialize_column_values_u128<F: Fn() -> I, I: Iterator<Item = u128>>(
+    iter_gen: F,
+    num_vals: u32,
+    output: &mut impl io::Write,
+) -> io::Result<()> {
+    let header = U128Header {
+        num_vals,
+        codec_type: U128FastFieldCodecType::CompactSpace,
+    };
+    header.serialize(output)?;
+    let compressor = CompactSpaceCompressor::train_from(iter_gen(), num_vals);
+    compressor.compress_into(iter_gen(), output)?;
 
-//     let null_index_footer = ColumnFooter {
-//         cardinality: value_index.get_cardinality(),
-//         null_index_codec: NullIndexCodec::Full,
-//         null_index_byte_range: 0..0,
-//     };
-//     append_null_index_footer(output, null_index_footer)?;
-//     append_format_version(output)?;
-
-//     Ok(())
-// }
+    Ok(())
+}
 
 /// Serializes the column with the codec with the best estimate on the data.
 pub fn serialize_column_values<T: MonotonicallyMappableToU64>(
@@ -279,20 +247,29 @@ pub(crate) fn serialize_given_codec(
     Ok(())
 }
 
-/// Helper function to serialize a column (autodetect from all codecs) and then open it
-pub fn serialize_and_load<T: MonotonicallyMappableToU64 + Ord + Default>(
-    column: &[T],
-) -> Arc<dyn ColumnValues<T>> {
-    let mut buffer = Vec::new();
-    super::serialize_column_values(&VecColumn::from(&column), &ALL_CODEC_TYPES, &mut buffer)
-        .unwrap();
-    super::open_u64_mapped(OwnedBytes::new(buffer)).unwrap()
-}
-
 #[cfg(test)]
-mod tests {
-    use super::*;
+pub mod tests {
+    use std::sync::Arc;
 
+    use common::OwnedBytes;
+
+    use super::*;
+    use crate::column_values::{open_u64_mapped, VecColumn};
+
+    const ALL_CODEC_TYPES: [FastFieldCodecType; 3] = [
+        FastFieldCodecType::Bitpacked,
+        FastFieldCodecType::Linear,
+        FastFieldCodecType::BlockwiseLinear,
+    ];
+
+    /// Helper function to serialize a column (autodetect from all codecs) and then open it
+    pub fn serialize_and_load<T: MonotonicallyMappableToU64 + Ord + Default>(
+        column: &[T],
+    ) -> Arc<dyn ColumnValues<T>> {
+        let mut buffer = Vec::new();
+        serialize_column_values(&VecColumn::from(&column), &ALL_CODEC_TYPES, &mut buffer).unwrap();
+        open_u64_mapped(OwnedBytes::new(buffer)).unwrap()
+    }
     #[test]
     fn test_serialize_deserialize_u128_header() {
         let original = U128Header {
@@ -319,7 +296,7 @@ mod tests {
         serialize_column_values(&col, &ALL_CODEC_TYPES, &mut buffer).unwrap();
         // TODO put the header as a footer so that it serves as a padding.
         // 5 bytes of header, 1 byte of value, 7 bytes of padding.
-        assert_eq!(buffer.len(), 5 + 1 + 7);
+        assert_eq!(buffer.len(), 5 + 1);
     }
 
     #[test]
@@ -328,7 +305,7 @@ mod tests {
         let col = VecColumn::from(&[true][..]);
         serialize_column_values(&col, &ALL_CODEC_TYPES, &mut buffer).unwrap();
         // 5 bytes of header, 0 bytes of value, 7 bytes of padding.
-        assert_eq!(buffer.len(), 5 + 7);
+        assert_eq!(buffer.len(), 5);
     }
 
     #[test]
@@ -338,6 +315,6 @@ mod tests {
         let col = VecColumn::from(&vals[..]);
         serialize_column_values(&col, &[FastFieldCodecType::Bitpacked], &mut buffer).unwrap();
         // Values are stored over 3 bits.
-        assert_eq!(buffer.len(), 7 + (3 * 80 / 8) + 7);
+        assert_eq!(buffer.len(), 7 + (3 * 80 / 8));
     }
 }

columnar/src/columnar/column_type.rs

@@ -1,98 +1,103 @@
-use crate::utils::{place_bits, select_bits};
+use std::net::Ipv6Addr;
+
 use crate::value::NumericalType;
 use crate::InvalidData;
 
-/// The column type represents the column type and can fit on 6-bits.
-///
-/// - bits[0..3]: Column category type.
-/// - bits[3..6]: Numerical type if necessary.
-#[derive(Hash, Eq, PartialEq, Debug, Clone, Copy)]
+/// The column type represents the column type.
+/// Any changes need to be propagated to `COLUMN_TYPES`.
+#[derive(Hash, Eq, PartialEq, Debug, Clone, Copy, Ord, PartialOrd)]
+#[repr(u8)]
 pub enum ColumnType {
-    Str,
-    Numerical(NumericalType),
-    Bool,
-    DateTime,
+    I64 = 0u8,
+    U64 = 1u8,
+    F64 = 2u8,
+    Bytes = 3u8,
+    Str = 4u8,
+    Bool = 5u8,
+    IpAddr = 6u8,
+    DateTime = 7u8,
 }
 
+// The order needs to match _exactly_ the order in the enum
+const COLUMN_TYPES: [ColumnType; 8] = [
+    ColumnType::I64,
+    ColumnType::U64,
+    ColumnType::F64,
+    ColumnType::Bytes,
+    ColumnType::Str,
+    ColumnType::Bool,
+    ColumnType::IpAddr,
+    ColumnType::DateTime,
+];
+
 impl ColumnType {
-    /// Encoded over 6 bits.
-    pub(crate) fn to_code(self) -> u8 {
-        let column_type_category;
-        let numerical_type_code: u8;
-        match self {
-            ColumnType::Str => {
-                column_type_category = ColumnTypeCategory::Str;
-                numerical_type_code = 0u8;
-            }
-            ColumnType::Numerical(numerical_type) => {
-                column_type_category = ColumnTypeCategory::Numerical;
-                numerical_type_code = numerical_type.to_code();
-            }
-            ColumnType::Bool => {
-                column_type_category = ColumnTypeCategory::Bool;
-                numerical_type_code = 0u8;
-            }
-            ColumnType::DateTime => {
-                column_type_category = ColumnTypeCategory::DateTime;
-                numerical_type_code = 0u8;
-            }
-        }
-        place_bits::<0, 3>(column_type_category.to_code()) | place_bits::<3, 6>(numerical_type_code)
+    pub fn to_code(self) -> u8 {
+        self as u8
     }
 
     pub(crate) fn try_from_code(code: u8) -> Result<ColumnType, InvalidData> {
-        if select_bits::<6, 8>(code) != 0u8 {
-            return Err(InvalidData);
-        }
-        let column_type_category_code = select_bits::<0, 3>(code);
-        let numerical_type_code = select_bits::<3, 6>(code);
-        let column_type_category = ColumnTypeCategory::try_from_code(column_type_category_code)?;
-        match column_type_category {
-            ColumnTypeCategory::Bool => {
-                if numerical_type_code != 0u8 {
-                    return Err(InvalidData);
-                }
-                Ok(ColumnType::Bool)
-            }
-            ColumnTypeCategory::Str => {
-                if numerical_type_code != 0u8 {
-                    return Err(InvalidData);
-                }
-                Ok(ColumnType::Str)
-            }
-            ColumnTypeCategory::Numerical => {
-                let numerical_type = NumericalType::try_from_code(numerical_type_code)?;
-                Ok(ColumnType::Numerical(numerical_type))
-            }
-            ColumnTypeCategory::DateTime => {
-                if numerical_type_code != 0u8 {
-                    return Err(InvalidData);
-                }
-                Ok(ColumnType::DateTime)
-            }
+        COLUMN_TYPES.get(code as usize).copied().ok_or(InvalidData)
+    }
+}
+
+impl From<NumericalType> for ColumnType {
+    fn from(numerical_type: NumericalType) -> Self {
+        match numerical_type {
+            NumericalType::I64 => ColumnType::I64,
+            NumericalType::U64 => ColumnType::U64,
+            NumericalType::F64 => ColumnType::F64,
         }
     }
 }
 
+impl ColumnType {
+    pub fn numerical_type(&self) -> Option<NumericalType> {
+        match self {
+            ColumnType::I64 => Some(NumericalType::I64),
+            ColumnType::U64 => Some(NumericalType::U64),
+            ColumnType::F64 => Some(NumericalType::F64),
+            ColumnType::Bytes
+            | ColumnType::Str
+            | ColumnType::Bool
+            | ColumnType::IpAddr
+            | ColumnType::DateTime => None,
+        }
+    }
+}
+
+// TODO remove if possible
 pub trait HasAssociatedColumnType: 'static + Send + Sync + Copy + PartialOrd {
     fn column_type() -> ColumnType;
+    fn default_value() -> Self;
 }
 
 impl HasAssociatedColumnType for u64 {
     fn column_type() -> ColumnType {
-        ColumnType::Numerical(NumericalType::U64)
+        ColumnType::U64
+    }
+
+    fn default_value() -> Self {
+        0u64
     }
 }
 
 impl HasAssociatedColumnType for i64 {
     fn column_type() -> ColumnType {
-        ColumnType::Numerical(NumericalType::I64)
+        ColumnType::I64
+    }
+
+    fn default_value() -> Self {
+        0i64
     }
 }
 
 impl HasAssociatedColumnType for f64 {
     fn column_type() -> ColumnType {
-        ColumnType::Numerical(NumericalType::F64)
+        ColumnType::F64
+    }
+
+    fn default_value() -> Self {
+        Default::default()
     }
 }
 
@@ -100,63 +105,45 @@ impl HasAssociatedColumnType for bool {
     fn column_type() -> ColumnType {
         ColumnType::Bool
     }
+    fn default_value() -> Self {
+        Default::default()
+    }
 }
 
 impl HasAssociatedColumnType for crate::DateTime {
     fn column_type() -> ColumnType {
         ColumnType::DateTime
     }
+    fn default_value() -> Self {
+        Default::default()
+    }
 }
 
-/// Column types are grouped into different categories that
-/// corresponds to the different types of `JsonValue` types.
-///
-/// The columnar writer will apply coercion rules to make sure that
-/// at most one column exist per `ColumnTypeCategory`.
-///
-/// See also [README.md].
-#[derive(Copy, Clone, Ord, PartialOrd, Eq, PartialEq, Debug)]
-#[repr(u8)]
-pub(crate) enum ColumnTypeCategory {
-    Bool = 0u8,
-    Str = 1u8,
-    Numerical = 2u8,
-    DateTime = 3u8,
-}
-
-impl ColumnTypeCategory {
-    pub fn to_code(self) -> u8 {
-        self as u8
+impl HasAssociatedColumnType for Ipv6Addr {
+    fn column_type() -> ColumnType {
+        ColumnType::IpAddr
     }
 
-    pub fn try_from_code(code: u8) -> Result<Self, InvalidData> {
-        match code {
-            0u8 => Ok(Self::Bool),
-            1u8 => Ok(Self::Str),
-            2u8 => Ok(Self::Numerical),
-            3u8 => Ok(Self::DateTime),
-            _ => Err(InvalidData),
-        }
+    fn default_value() -> Self {
+        Ipv6Addr::from([0u8; 16])
     }
 }
 
 #[cfg(test)]
 mod tests {
-    use std::collections::HashSet;
-
     use super::*;
     use crate::Cardinality;
 
     #[test]
     fn test_column_type_to_code() {
-        let mut column_type_set: HashSet<ColumnType> = HashSet::new();
-        for code in u8::MIN..=u8::MAX {
-            if let Ok(column_type) = ColumnType::try_from_code(code) {
-                assert_eq!(column_type.to_code(), code);
-                assert!(column_type_set.insert(column_type));
+        for (code, expected_column_type) in super::COLUMN_TYPES.iter().copied().enumerate() {
+            if let Ok(column_type) = ColumnType::try_from_code(code as u8) {
+                assert_eq!(column_type, expected_column_type);
             }
         }
-        assert_eq!(column_type_set.len(), 3 + 3);
+        for code in COLUMN_TYPES.len() as u8..=u8::MAX {
+            assert!(ColumnType::try_from_code(code as u8).is_err());
+        }
     }
 
     #[test]

columnar/src/columnar/merge.rs

@@ -0,0 +1,248 @@
+use std::collections::HashMap;
+use std::io;
+
+use super::writer::ColumnarSerializer;
+use crate::columnar::ColumnarReader;
+use crate::dynamic_column::DynamicColumn;
+use crate::{Cardinality, ColumnType};
+
+pub enum MergeDocOrder {
+    /// Columnar tables are simply stacked one above the other.
+    /// If the i-th columnar_readers has n_rows_i rows, then
+    /// in the resulting columnar,
+    /// rows [r0..n_row_0) contains the row of columnar_readers[0], in ordder
+    /// rows [n_row_0..n_row_0 + n_row_1 contains the row of columnar_readers[1], in order.
+    /// ..
+    Stack,
+    /// Some more complex mapping, that can interleaves rows from the different readers and
+    /// possibly drop rows.
+    Complex(()),
+}
+
+pub fn merge_columnar(
+    columnar_readers: &[ColumnarReader],
+    mapping: MergeDocOrder,
+    output: &mut impl io::Write,
+) -> io::Result<()> {
+    let mut serializer = ColumnarSerializer::new(output);
+
+    // TODO handle dictionary merge for Str/Bytes column
+    let field_name_to_group = group_columns_for_merge(columnar_readers)?;
+    for (column_name, category_to_columns) in field_name_to_group {
+        for (_category, columns_to_merge) in category_to_columns {
+            let column_type = columns_to_merge[0].column_type();
+            let mut column_serialzier =
+                serializer.serialize_column(column_name.as_bytes(), column_type);
+            merge_columns(
+                column_type,
+                &columns_to_merge,
+                &mapping,
+                &mut column_serialzier,
+            )?;
+        }
+    }
+    serializer.finalize()?;
+
+    Ok(())
+}
+
+/// Column types are grouped into different categories.
+/// After merge, all columns belonging to the same category are coerced to
+/// the same column type.
+///
+/// In practise, today, only Numerical colummns are coerced into one type today.
+///
+/// See also [README.md].
+#[derive(Copy, Clone, Eq, PartialEq, Hash, Debug)]
+#[repr(u8)]
+pub enum ColumnTypeCategory {
+    Bool,
+    Str,
+    Numerical,
+    DateTime,
+    Bytes,
+    IpAddr,
+}
+
+impl From<ColumnType> for ColumnTypeCategory {
+    fn from(column_type: ColumnType) -> Self {
+        match column_type {
+            ColumnType::I64 => ColumnTypeCategory::Numerical,
+            ColumnType::U64 => ColumnTypeCategory::Numerical,
+            ColumnType::F64 => ColumnTypeCategory::Numerical,
+            ColumnType::Bytes => ColumnTypeCategory::Bytes,
+            ColumnType::Str => ColumnTypeCategory::Str,
+            ColumnType::Bool => ColumnTypeCategory::Bool,
+            ColumnType::IpAddr => ColumnTypeCategory::IpAddr,
+            ColumnType::DateTime => ColumnTypeCategory::DateTime,
+        }
+    }
+}
+
+pub fn detect_cardinality(columns: &[DynamicColumn]) -> Cardinality {
+    if columns
+        .iter()
+        .any(|column| column.get_cardinality().is_multivalue())
+    {
+        return Cardinality::Multivalued;
+    }
+    if columns
+        .iter()
+        .any(|column| column.get_cardinality().is_optional())
+    {
+        return Cardinality::Optional;
+    }
+    Cardinality::Full
+}
+
+pub fn compute_num_docs(columns: &[DynamicColumn], mapping: &MergeDocOrder) -> usize {
+    // TODO handle deletes
+
+    0
+}
+
+pub fn merge_columns(
+    column_type: ColumnType,
+    columns: &[DynamicColumn],
+    mapping: &MergeDocOrder,
+    column_serializer: &mut impl io::Write,
+) -> io::Result<()> {
+    let cardinality = detect_cardinality(columns);
+
+    Ok(())
+}
+
+pub fn group_columns_for_merge(
+    columnar_readers: &[ColumnarReader],
+) -> io::Result<HashMap<String, HashMap<ColumnTypeCategory, Vec<DynamicColumn>>>> {
+    // Each column name may have multiple types of column associated.
+    // For merging we are interested in the same column type category since they can be merged.
+    let mut field_name_to_group: HashMap<String, HashMap<ColumnTypeCategory, Vec<DynamicColumn>>> =
+        HashMap::new();
+
+    for columnar_reader in columnar_readers {
+        let column_name_and_handle = columnar_reader.list_columns()?;
+        for (column_name, handle) in column_name_and_handle {
+            let column_type_to_handles = field_name_to_group
+                .entry(column_name.to_string())
+                .or_default();
+
+            let columns = column_type_to_handles
+                .entry(handle.column_type().into())
+                .or_default();
+            columns.push(handle.open()?);
+        }
+    }
+
+    normalize_columns(&mut field_name_to_group);
+
+    Ok(field_name_to_group)
+}
+
+/// Coerce numerical type columns to the same type
+/// TODO rename to `coerce_columns`
+fn normalize_columns(map: &mut HashMap<String, HashMap<ColumnTypeCategory, Vec<DynamicColumn>>>) {
+    for (_field_name, type_category_to_columns) in map.iter_mut() {
+        for (type_category, columns) in type_category_to_columns {
+            if type_category == &ColumnTypeCategory::Numerical {
+                let casted_columns = cast_to_common_numerical_column(&columns);
+                *columns = casted_columns;
+            }
+        }
+    }
+}
+
+/// Receives a list of columns of numerical types (u64, i64, f64)
+///
+/// Returns a list of `DynamicColumn` which are all of the same numerical type
+fn cast_to_common_numerical_column(columns: &[DynamicColumn]) -> Vec<DynamicColumn> {
+    assert!(columns
+        .iter()
+        .all(|column| column.column_type().numerical_type().is_some()));
+    let coerce_to_i64: Vec<_> = columns
+        .iter()
+        .filter_map(|column| column.clone().coerce_to_i64())
+        .collect();
+
+    if coerce_to_i64.len() == columns.len() {
+        return coerce_to_i64;
+    }
+
+    let coerce_to_u64: Vec<_> = columns
+        .iter()
+        .filter_map(|column| column.clone().coerce_to_u64())
+        .collect();
+
+    if coerce_to_u64.len() == columns.len() {
+        return coerce_to_u64;
+    }
+
+    columns
+        .iter()
+        .map(|column| {
+            column
+                .clone()
+                .coerce_to_f64()
+                .expect("couldn't cast column to f64")
+        })
+        .collect()
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::ColumnarWriter;
+
+    #[test]
+    fn test_column_coercion() {
+        // i64 type
+        let columnar1 = {
+            let mut dataframe_writer = ColumnarWriter::default();
+            dataframe_writer.record_numerical(1u32, "numbers", 1i64);
+            let mut buffer: Vec<u8> = Vec::new();
+            dataframe_writer.serialize(2, &mut buffer).unwrap();
+            ColumnarReader::open(buffer).unwrap()
+        };
+        // u64 type
+        let columnar2 = {
+            let mut dataframe_writer = ColumnarWriter::default();
+            dataframe_writer.record_numerical(1u32, "numbers", u64::MAX - 100);
+            let mut buffer: Vec<u8> = Vec::new();
+            dataframe_writer.serialize(2, &mut buffer).unwrap();
+            ColumnarReader::open(buffer).unwrap()
+        };
+
+        // f64 type
+        let columnar3 = {
+            let mut dataframe_writer = ColumnarWriter::default();
+            dataframe_writer.record_numerical(1u32, "numbers", 30.5);
+            let mut buffer: Vec<u8> = Vec::new();
+            dataframe_writer.serialize(2, &mut buffer).unwrap();
+            ColumnarReader::open(buffer).unwrap()
+        };
+
+        let column_map =
+            group_columns_for_merge(&[columnar1.clone(), columnar2.clone(), columnar3.clone()])
+                .unwrap();
+        assert_eq!(column_map.len(), 1);
+        let cat_to_columns = column_map.get("numbers").unwrap();
+        assert_eq!(cat_to_columns.len(), 1);
+
+        let numerical = cat_to_columns.get(&ColumnTypeCategory::Numerical).unwrap();
+        assert!(numerical.iter().all(|column| column.is_f64()));
+
+        let column_map = group_columns_for_merge(&[columnar1.clone(), columnar1.clone()]).unwrap();
+        assert_eq!(column_map.len(), 1);
+        let cat_to_columns = column_map.get("numbers").unwrap();
+        assert_eq!(cat_to_columns.len(), 1);
+        let numerical = cat_to_columns.get(&ColumnTypeCategory::Numerical).unwrap();
+        assert!(numerical.iter().all(|column| column.is_i64()));
+
+        let column_map = group_columns_for_merge(&[columnar2.clone(), columnar2.clone()]).unwrap();
+        assert_eq!(column_map.len(), 1);
+        let cat_to_columns = column_map.get("numbers").unwrap();
+        assert_eq!(cat_to_columns.len(), 1);
+        let numerical = cat_to_columns.get(&ColumnTypeCategory::Numerical).unwrap();
+        assert!(numerical.iter().all(|column| column.is_u64()));
+    }
+}

columnar/src/columnar/merge_index.rs

@@ -0,0 +1 @@
+

columnar/src/columnar/mod.rs

@@ -1,28 +1,11 @@
-// Copyright (C) 2022 Quickwit, Inc.
-//
-// Quickwit is offered under the AGPL v3.0 and as commercial software.
-// For commercial licensing, contact us at hello@quickwit.io.
-//
-// AGPL:
-// This program is free software: you can redistribute it and/or modify
-// it under the terms of the GNU Affero General Public License as
-// published by the Free Software Foundation, either version 3 of the
-// License, or (at your option) any later version.
-//
-// This program is distributed in the hope that it will be useful,
-// but WITHOUT ANY WARRANTY; without even the implied warranty of
-// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-// GNU Affero General Public License for more details.
-//
-// You should have received a copy of the GNU Affero General Public License
-// along with this program. If not, see <http://www.gnu.org/licenses/>.
-//
-
 mod column_type;
 mod format_version;
+mod merge;
+mod merge_index;
 mod reader;
 mod writer;
 
 pub use column_type::{ColumnType, HasAssociatedColumnType};
+pub use merge::{merge_columnar, MergeDocOrder};
 pub use reader::ColumnarReader;
 pub use writer::ColumnarWriter;

columnar/src/columnar/reader/mod.rs

@@ -13,6 +13,7 @@ fn io_invalid_data(msg: String) -> io::Error {
 
 /// The ColumnarReader makes it possible to access a set of columns
 /// associated to field names.
+#[derive(Clone)]
 pub struct ColumnarReader {
     column_dictionary: Dictionary<RangeSSTable>,
     column_data: FileSlice,

columnar/src/columnar/writer/column_operation.rs

@@ -1,3 +1,5 @@
+use std::net::Ipv6Addr;
+
 use crate::dictionary::UnorderedId;
 use crate::utils::{place_bits, pop_first_byte, select_bits};
 use crate::value::NumericalValue;
@@ -25,12 +27,12 @@ struct ColumnOperationMetadata {
 
 impl ColumnOperationMetadata {
     fn to_code(self) -> u8 {
-        place_bits::<0, 4>(self.len) | place_bits::<4, 8>(self.op_type.to_code())
+        place_bits::<0, 6>(self.len) | place_bits::<6, 8>(self.op_type.to_code())
     }
 
     fn try_from_code(code: u8) -> Result<Self, InvalidData> {
-        let len = select_bits::<0, 4>(code);
-        let typ_code = select_bits::<4, 8>(code);
+        let len = select_bits::<0, 6>(code);
+        let typ_code = select_bits::<6, 8>(code);
         let column_type = ColumnOperationType::try_from_code(typ_code)?;
         Ok(ColumnOperationMetadata {
             op_type: column_type,
@@ -142,9 +144,21 @@ impl SymbolValue for bool {
     }
 }
 
+impl SymbolValue for Ipv6Addr {
+    fn serialize(self, buffer: &mut [u8]) -> u8 {
+        buffer[0..16].copy_from_slice(&self.octets());
+        16
+    }
+
+    fn deserialize(bytes: &[u8]) -> Self {
+        let octets: [u8; 16] = bytes[0..16].try_into().unwrap();
+        Ipv6Addr::from(octets)
+    }
+}
+
 #[derive(Default)]
 struct MiniBuffer {
-    pub bytes: [u8; 10],
+    pub bytes: [u8; 17],
     pub len: u8,
 }
 

columnar/src/columnar/writer/column_writers.rs

@@ -184,10 +184,12 @@ impl CompatibleNumericalTypes {
 }
 
 impl NumericalColumnWriter {
-    pub fn column_type_and_cardinality(&self, num_docs: RowId) -> (NumericalType, Cardinality) {
-        let numerical_type = self.compatible_numerical_types.to_numerical_type();
-        let cardinality = self.column_writer.get_cardinality(num_docs);
-        (numerical_type, cardinality)
+    pub fn numerical_type(&self) -> NumericalType {
+        self.compatible_numerical_types.to_numerical_type()
+    }
+
+    pub fn cardinality(&self, num_docs: RowId) -> Cardinality {
+        self.column_writer.get_cardinality(num_docs)
     }
 
     pub fn record_numerical_value(
@@ -209,15 +211,15 @@ impl NumericalColumnWriter {
     }
 }
 
-#[derive(Copy, Clone, Default)]
-pub(crate) struct StrColumnWriter {
+#[derive(Copy, Clone)]
+pub(crate) struct StrOrBytesColumnWriter {
     pub(crate) dictionary_id: u32,
     pub(crate) column_writer: ColumnWriter,
 }
 
-impl StrColumnWriter {
-    pub(crate) fn with_dictionary_id(dictionary_id: u32) -> StrColumnWriter {
-        StrColumnWriter {
+impl StrOrBytesColumnWriter {
+    pub(crate) fn with_dictionary_id(dictionary_id: u32) -> StrOrBytesColumnWriter {
+        StrOrBytesColumnWriter {
             dictionary_id,
             column_writer: Default::default(),
         }

columnar/src/columnar/writer/mod.rs

@@ -4,17 +4,20 @@ mod serializer;
 mod value_index;
 
 use std::io;
+use std::net::Ipv6Addr;
 
 use column_operation::ColumnOperation;
 use common::CountingWriter;
-use serializer::ColumnarSerializer;
+pub(crate) use serializer::ColumnarSerializer;
 use stacker::{Addr, ArenaHashMap, MemoryArena};
 
 use crate::column_index::SerializableColumnIndex;
-use crate::column_values::{ColumnValues, MonotonicallyMappableToU64, VecColumn};
-use crate::columnar::column_type::{ColumnType, ColumnTypeCategory};
+use crate::column_values::{
+    ColumnValues, MonotonicallyMappableToU128, MonotonicallyMappableToU64, VecColumn,
+};
+use crate::columnar::column_type::ColumnType;
 use crate::columnar::writer::column_writers::{
-    ColumnWriter, NumericalColumnWriter, StrColumnWriter,
+    ColumnWriter, NumericalColumnWriter, StrOrBytesColumnWriter,
 };
 use crate::columnar::writer::value_index::{IndexBuilder, PreallocatedIndexBuilders};
 use crate::dictionary::{DictionaryBuilder, TermIdMapping, UnorderedId};
@@ -30,6 +33,7 @@ struct SpareBuffers {
     u64_values: Vec<u64>,
     f64_values: Vec<f64>,
     bool_values: Vec<bool>,
+    ip_addr_values: Vec<Ipv6Addr>,
 }
 
 /// Makes it possible to create a new columnar.
@@ -49,7 +53,9 @@ pub struct ColumnarWriter {
     numerical_field_hash_map: ArenaHashMap,
     datetime_field_hash_map: ArenaHashMap,
     bool_field_hash_map: ArenaHashMap,
+    ip_addr_field_hash_map: ArenaHashMap,
     bytes_field_hash_map: ArenaHashMap,
+    str_field_hash_map: ArenaHashMap,
     arena: MemoryArena,
     // Dictionaries used to store dictionary-encoded values.
     dictionaries: Vec<DictionaryBuilder>,
@@ -61,7 +67,9 @@ impl Default for ColumnarWriter {
         ColumnarWriter {
             numerical_field_hash_map: ArenaHashMap::new(10_000),
             bool_field_hash_map: ArenaHashMap::new(10_000),
+            ip_addr_field_hash_map: ArenaHashMap::new(10_000),
             bytes_field_hash_map: ArenaHashMap::new(10_000),
+            str_field_hash_map: ArenaHashMap::new(10_000),
             datetime_field_hash_map: ArenaHashMap::new(10_000),
             dictionaries: Vec::new(),
             arena: MemoryArena::default(),
@@ -93,15 +101,34 @@ impl ColumnarWriter {
             + self.numerical_field_hash_map.mem_usage()
             + self.bool_field_hash_map.mem_usage()
             + self.bytes_field_hash_map.mem_usage()
+            + self.str_field_hash_map.mem_usage()
+            + self.ip_addr_field_hash_map.mem_usage()
+            + self.datetime_field_hash_map.mem_usage()
     }
 
     pub fn record_column_type(&mut self, column_name: &str, column_type: ColumnType) {
         match column_type {
-            ColumnType::Str => {
+            ColumnType::Str | ColumnType::Bytes => {
+                let (hash_map, dictionaries) = (
+                    if column_type == ColumnType::Str {
+                        &mut self.str_field_hash_map
+                    } else {
+                        &mut self.bytes_field_hash_map
+                    },
+                    &mut self.dictionaries,
+                );
                 mutate_or_create_column(
-                    &mut self.bytes_field_hash_map,
+                    hash_map,
                     column_name,
-                    |column_opt: Option<StrColumnWriter>| column_opt.unwrap_or_default(),
+                    |column_opt: Option<StrOrBytesColumnWriter>| {
+                        if let Some(column_writer) = column_opt {
+                            column_writer
+                        } else {
+                            let dictionary_id = dictionaries.len() as u32;
+                            dictionaries.push(DictionaryBuilder::default());
+                            StrOrBytesColumnWriter::with_dictionary_id(dictionary_id)
+                        }
+                    },
                 );
             }
             ColumnType::Bool => {
@@ -118,7 +145,8 @@ impl ColumnarWriter {
                     |column_opt: Option<ColumnWriter>| column_opt.unwrap_or_default(),
                 );
             }
-            ColumnType::Numerical(numerical_type) => {
+            ColumnType::I64 | ColumnType::F64 | ColumnType::U64 => {
+                let numerical_type = column_type.numerical_type().unwrap();
                 mutate_or_create_column(
                     &mut self.numerical_field_hash_map,
                     column_name,
@@ -129,6 +157,11 @@ impl ColumnarWriter {
                     },
                 );
             }
+            ColumnType::IpAddr => mutate_or_create_column(
+                &mut self.ip_addr_field_hash_map,
+                column_name,
+                |column_opt: Option<ColumnWriter>| column_opt.unwrap_or_default(),
+            ),
         }
     }
 
@@ -162,6 +195,22 @@ impl ColumnarWriter {
         );
     }
 
+    pub fn record_ip_addr(&mut self, doc: RowId, column_name: &str, ip_addr: Ipv6Addr) {
+        assert!(
+            !column_name.as_bytes().contains(&0u8),
+            "key may not contain the 0 byte"
+        );
+        let (hash_map, arena) = (&mut self.ip_addr_field_hash_map, &mut self.arena);
+        hash_map.mutate_or_create(
+            column_name.as_bytes(),
+            |column_opt: Option<ColumnWriter>| {
+                let mut column: ColumnWriter = column_opt.unwrap_or_default();
+                column.record(doc, ip_addr, arena);
+                column
+            },
+        );
+    }
+
     pub fn record_bool(&mut self, doc: RowId, column_name: &str, val: bool) {
         let (hash_map, arena) = (&mut self.bool_field_hash_map, &mut self.arena);
         mutate_or_create_column(hash_map, column_name, |column_opt: Option<ColumnWriter>| {
@@ -182,19 +231,18 @@ impl ColumnarWriter {
 
     pub fn record_str(&mut self, doc: RowId, column_name: &str, value: &str) {
         let (hash_map, arena, dictionaries) = (
-            &mut self.bytes_field_hash_map,
+            &mut self.str_field_hash_map,
             &mut self.arena,
             &mut self.dictionaries,
         );
-        mutate_or_create_column(
-            hash_map,
-            column_name,
-            |column_opt: Option<StrColumnWriter>| {
-                let mut column: StrColumnWriter = column_opt.unwrap_or_else(|| {
+        hash_map.mutate_or_create(
+            column_name.as_bytes(),
+            |column_opt: Option<StrOrBytesColumnWriter>| {
+                let mut column: StrOrBytesColumnWriter = column_opt.unwrap_or_else(|| {
                     // Each column has its own dictionary
                     let dictionary_id = dictionaries.len() as u32;
                     dictionaries.push(DictionaryBuilder::default());
-                    StrColumnWriter::with_dictionary_id(dictionary_id)
+                    StrOrBytesColumnWriter::with_dictionary_id(dictionary_id)
                 });
                 column.record_bytes(doc, value.as_bytes(), dictionaries, arena);
                 column
@@ -202,35 +250,79 @@ impl ColumnarWriter {
         );
     }
 
+    pub fn record_bytes(&mut self, doc: RowId, column_name: &str, value: &[u8]) {
+        assert!(
+            !column_name.as_bytes().contains(&0u8),
+            "key may not contain the 0 byte"
+        );
+        let (hash_map, arena, dictionaries) = (
+            &mut self.bytes_field_hash_map,
+            &mut self.arena,
+            &mut self.dictionaries,
+        );
+        hash_map.mutate_or_create(
+            column_name.as_bytes(),
+            |column_opt: Option<StrOrBytesColumnWriter>| {
+                let mut column: StrOrBytesColumnWriter = column_opt.unwrap_or_else(|| {
+                    // Each column has its own dictionary
+                    let dictionary_id = dictionaries.len() as u32;
+                    dictionaries.push(DictionaryBuilder::default());
+                    StrOrBytesColumnWriter::with_dictionary_id(dictionary_id)
+                });
+                column.record_bytes(doc, value, dictionaries, arena);
+                column
+            },
+        );
+    }
     pub fn serialize(&mut self, num_docs: RowId, wrt: &mut dyn io::Write) -> io::Result<()> {
         let mut serializer = ColumnarSerializer::new(wrt);
-        let mut columns: Vec<(&[u8], ColumnTypeCategory, Addr)> = self
+        let mut columns: Vec<(&[u8], ColumnType, Addr)> = self
             .numerical_field_hash_map
             .iter()
-            .map(|(column_name, addr, _)| (column_name, ColumnTypeCategory::Numerical, addr))
+            .map(|(column_name, addr, _)| {
+                let numerical_column_writer: NumericalColumnWriter =
+                    self.numerical_field_hash_map.read(addr);
+                let column_type = numerical_column_writer.numerical_type().into();
+                (column_name, column_type, addr)
+            })
             .collect();
         columns.extend(
             self.bytes_field_hash_map
                 .iter()
-                .map(|(column_name, addr, _)| (column_name, ColumnTypeCategory::Str, addr)),
+                .map(|(term, addr, _)| (term, ColumnType::Bytes, addr)),
+        );
+        columns.extend(
+            self.str_field_hash_map
+                .iter()
+                .map(|(column_name, addr, _)| (column_name, ColumnType::Str, addr)),
         );
         columns.extend(
             self.bool_field_hash_map
                 .iter()
-                .map(|(column_name, addr, _)| (column_name, ColumnTypeCategory::Bool, addr)),
+                .map(|(column_name, addr, _)| (column_name, ColumnType::Bool, addr)),
+        );
+        columns.extend(
+            self.ip_addr_field_hash_map
+                .iter()
+                .map(|(column_name, addr, _)| (column_name, ColumnType::IpAddr, addr)),
         );
         columns.extend(
             self.datetime_field_hash_map
                 .iter()
-                .map(|(column_name, addr, _)| (column_name, ColumnTypeCategory::DateTime, addr)),
+                .map(|(column_name, addr, _)| (column_name, ColumnType::DateTime, addr)),
         );
         columns.sort_unstable_by_key(|(column_name, col_type, _)| (*column_name, *col_type));
+
         let (arena, buffers, dictionaries) = (&self.arena, &mut self.buffers, &self.dictionaries);
         let mut symbol_byte_buffer: Vec<u8> = Vec::new();
-        for (column_name, bytes_or_numerical, addr) in columns {
-            match bytes_or_numerical {
-                ColumnTypeCategory::Bool => {
-                    let column_writer: ColumnWriter = self.bool_field_hash_map.read(addr);
+        for (column_name, column_type, addr) in columns {
+            match column_type {
+                ColumnType::Bool | ColumnType::DateTime => {
+                    let column_writer: ColumnWriter = if column_type == ColumnType::Bool {
+                        self.bool_field_hash_map.read(addr)
+                    } else {
+                        self.datetime_field_hash_map.read(addr)
+                    };
                     let cardinality = column_writer.get_cardinality(num_docs);
                     let mut column_serializer =
                         serializer.serialize_column(column_name, ColumnType::Bool);
@@ -242,29 +334,50 @@ impl ColumnarWriter {
                         &mut column_serializer,
                     )?;
                 }
-                ColumnTypeCategory::Str => {
-                    let str_column_writer: StrColumnWriter = self.bytes_field_hash_map.read(addr);
-                    let dictionary_builder =
-                        &dictionaries[str_column_writer.dictionary_id as usize];
-                    let cardinality = str_column_writer.column_writer.get_cardinality(num_docs);
+                ColumnType::IpAddr => {
+                    let column_writer: ColumnWriter = self.ip_addr_field_hash_map.read(addr);
+                    let cardinality = column_writer.get_cardinality(num_docs);
                     let mut column_serializer =
-                        serializer.serialize_column(column_name, ColumnType::Str);
-                    serialize_bytes_column(
+                        serializer.serialize_column(column_name, ColumnType::IpAddr);
+                    serialize_ip_addr_column(
                         cardinality,
                         num_docs,
-                        dictionary_builder,
-                        str_column_writer.operation_iterator(arena, &mut symbol_byte_buffer),
+                        column_writer.operation_iterator(arena, &mut symbol_byte_buffer),
                         buffers,
                         &mut column_serializer,
                     )?;
                 }
-                ColumnTypeCategory::Numerical => {
+                ColumnType::Bytes | ColumnType::Str => {
+                    let str_or_bytes_column_writer: StrOrBytesColumnWriter =
+                        if column_type == ColumnType::Bytes {
+                            self.bytes_field_hash_map.read(addr)
+                        } else {
+                            self.str_field_hash_map.read(addr)
+                        };
+                    let dictionary_builder =
+                        &dictionaries[str_or_bytes_column_writer.dictionary_id as usize];
+                    let cardinality = str_or_bytes_column_writer
+                        .column_writer
+                        .get_cardinality(num_docs);
+                    let mut column_serializer =
+                        serializer.serialize_column(column_name, column_type);
+                    serialize_bytes_or_str_column(
+                        cardinality,
+                        num_docs,
+                        dictionary_builder,
+                        str_or_bytes_column_writer
+                            .operation_iterator(arena, &mut symbol_byte_buffer),
+                        buffers,
+                        &mut column_serializer,
+                    )?;
+                }
+                ColumnType::I64 | ColumnType::F64 | ColumnType::U64 => {
                     let numerical_column_writer: NumericalColumnWriter =
                         self.numerical_field_hash_map.read(addr);
-                    let (numerical_type, cardinality) =
-                        numerical_column_writer.column_type_and_cardinality(num_docs);
-                    let mut column_serializer = serializer
-                        .serialize_column(column_name, ColumnType::Numerical(numerical_type));
+                    let numerical_type = column_type.numerical_type().unwrap();
+                    let cardinality = numerical_column_writer.cardinality(num_docs);
+                    let mut column_serializer =
+                        serializer.serialize_column(column_name, ColumnType::from(numerical_type));
                     serialize_numerical_column(
                         cardinality,
                         num_docs,
@@ -274,20 +387,6 @@ impl ColumnarWriter {
                         &mut column_serializer,
                     )?;
                 }
-                ColumnTypeCategory::DateTime => {
-                    let column_writer: ColumnWriter = self.datetime_field_hash_map.read(addr);
-                    let cardinality = column_writer.get_cardinality(num_docs);
-                    let mut column_serializer =
-                        serializer.serialize_column(column_name, ColumnType::DateTime);
-                    serialize_numerical_column(
-                        cardinality,
-                        num_docs,
-                        NumericalType::I64,
-                        column_writer.operation_iterator(arena, &mut symbol_byte_buffer),
-                        buffers,
-                        &mut column_serializer,
-                    )?;
-                }
             };
         }
         serializer.finalize()?;
@@ -295,7 +394,7 @@ impl ColumnarWriter {
     }
 }
 
-fn serialize_bytes_column(
+fn serialize_bytes_or_str_column(
     cardinality: Cardinality,
     num_docs: RowId,
     dictionary_builder: &DictionaryBuilder,
@@ -322,7 +421,7 @@ fn serialize_bytes_column(
             ColumnOperation::NewDoc(doc) => ColumnOperation::NewDoc(doc),
         }
     });
-    serialize_column(
+    send_to_serialize_column_mappable_to_u64(
         operation_iterator,
         cardinality,
         num_docs,
@@ -351,7 +450,7 @@ fn serialize_numerical_column(
     } = buffers;
     match numerical_type {
         NumericalType::I64 => {
-            serialize_column(
+            send_to_serialize_column_mappable_to_u64(
                 coerce_numerical_symbol::<i64>(op_iterator),
                 cardinality,
                 num_docs,
@@ -361,7 +460,7 @@ fn serialize_numerical_column(
             )?;
         }
         NumericalType::U64 => {
-            serialize_column(
+            send_to_serialize_column_mappable_to_u64(
                 coerce_numerical_symbol::<u64>(op_iterator),
                 cardinality,
                 num_docs,
@@ -371,7 +470,7 @@ fn serialize_numerical_column(
             )?;
         }
         NumericalType::F64 => {
-            serialize_column(
+            send_to_serialize_column_mappable_to_u64(
                 coerce_numerical_symbol::<f64>(op_iterator),
                 cardinality,
                 num_docs,
@@ -396,7 +495,7 @@ fn serialize_bool_column(
         bool_values,
         ..
     } = buffers;
-    serialize_column(
+    send_to_serialize_column_mappable_to_u64(
         column_operations_it,
         cardinality,
         num_docs,
@@ -407,7 +506,76 @@ fn serialize_bool_column(
     Ok(())
 }
 
-fn serialize_column<
+fn serialize_ip_addr_column(
+    cardinality: Cardinality,
+    num_docs: RowId,
+    column_operations_it: impl Iterator<Item = ColumnOperation<Ipv6Addr>>,
+    buffers: &mut SpareBuffers,
+    wrt: &mut impl io::Write,
+) -> io::Result<()> {
+    let SpareBuffers {
+        value_index_builders,
+        ip_addr_values,
+        ..
+    } = buffers;
+    send_to_serialize_column_mappable_to_u128(
+        column_operations_it,
+        cardinality,
+        num_docs,
+        value_index_builders,
+        ip_addr_values,
+        wrt,
+    )?;
+    Ok(())
+}
+
+fn send_to_serialize_column_mappable_to_u128<
+    T: Copy + std::fmt::Debug + Send + Sync + MonotonicallyMappableToU128 + PartialOrd,
+>(
+    op_iterator: impl Iterator<Item = ColumnOperation<T>>,
+    cardinality: Cardinality,
+    num_docs: RowId,
+    value_index_builders: &mut PreallocatedIndexBuilders,
+    values: &mut Vec<T>,
+    mut wrt: impl io::Write,
+) -> io::Result<()>
+where
+    for<'a> VecColumn<'a, T>: ColumnValues<T>,
+{
+    values.clear();
+    // TODO: split index and values
+    let serializable_column_index = match cardinality {
+        Cardinality::Full => {
+            consume_operation_iterator(
+                op_iterator,
+                value_index_builders.borrow_required_index_builder(),
+                values,
+            );
+            SerializableColumnIndex::Full
+        }
+        Cardinality::Optional => {
+            let optional_index_builder = value_index_builders.borrow_optional_index_builder();
+            consume_operation_iterator(op_iterator, optional_index_builder, values);
+            let optional_index = optional_index_builder.finish(num_docs);
+            SerializableColumnIndex::Optional(Box::new(optional_index))
+        }
+        Cardinality::Multivalued => {
+            let multivalued_index_builder = value_index_builders.borrow_multivalued_index_builder();
+            consume_operation_iterator(op_iterator, multivalued_index_builder, values);
+            let multivalued_index = multivalued_index_builder.finish(num_docs);
+            SerializableColumnIndex::Multivalued(Box::new(multivalued_index))
+        }
+    };
+    crate::column::serialize_column_mappable_to_u128(
+        serializable_column_index,
+        || values.iter().cloned(),
+        values.len() as u32,
+        &mut wrt,
+    )?;
+    Ok(())
+}
+
+fn send_to_serialize_column_mappable_to_u64<
     T: Copy + Default + std::fmt::Debug + Send + Sync + MonotonicallyMappableToU64 + PartialOrd,
 >(
     op_iterator: impl Iterator<Item = ColumnOperation<T>>,
@@ -440,11 +608,10 @@ where
             let multivalued_index_builder = value_index_builders.borrow_multivalued_index_builder();
             consume_operation_iterator(op_iterator, multivalued_index_builder, values);
             let multivalued_index = multivalued_index_builder.finish(num_docs);
-            todo!();
-            // SerializableColumnIndex::Multivalued(Box::new(multivalued_index))
+            SerializableColumnIndex::Multivalued(Box::new(multivalued_index))
         }
     };
-    crate::column::serialize_column_u64(
+    crate::column::serialize_column_mappable_to_u64(
         serializable_column_index,
         &VecColumn::from(&values[..]),
         &mut wrt,
@@ -482,59 +649,12 @@ fn consume_operation_iterator<T: std::fmt::Debug, TIndexBuilder: IndexBuilder>(
     }
 }
 
-// /// Serializes the column with the codec with the best estimate on the data.
-// fn serialize_numerical<T: MonotonicallyMappableToU64>(
-//     value_index: ValueIndexInfo,
-//     typed_column: impl Column<T>,
-//     output: &mut impl io::Write,
-//     codecs: &[FastFieldCodecType],
-// ) -> io::Result<()> {
-
-//     let counting_writer = CountingWriter::wrap(output);
-//     serialize_value_index(value_index, output)?;
-//     let value_index_len = counting_writer.written_bytes();
-//     let output = counting_writer.finish();
-
-//     serialize_column(value_index, output)?;
-//     let column = monotonic_map_column(
-//         typed_column,
-//         crate::column::monotonic_mapping::StrictlyMonotonicMappingToInternal::<T>::new(),
-//     );
-//     let header = Header::compute_header(&column, codecs).ok_or_else(|| {
-//         io::Error::new(
-//             io::ErrorKind::InvalidInput,
-//             format!(
-//                 "Data cannot be serialized with this list of codec. {:?}",
-//                 codecs
-//             ),
-//         )
-//     })?;
-//     header.serialize(output)?;
-//     let normalized_column = header.normalize_column(column);
-//     assert_eq!(normalized_column.min_value(), 0u64);
-//     serialize_given_codec(normalized_column, header.codec_type, output)?;
-
-//     let column_header = ColumnFooter {
-//         value_index_len: todo!(),
-//         cardinality: todo!(),
-//     };
-
-//     let null_index_footer = NullIndexFooter {
-//         cardinality: value_index.get_cardinality(),
-//         null_index_codec: NullIndexCodec::Full,
-//         null_index_byte_range: 0..0,
-//     };
-//     append_null_index_footer(output, null_index_footer)?;
-//     Ok(())
-// }
-
 #[cfg(test)]
 mod tests {
-    use column_operation::ColumnOperation;
     use stacker::MemoryArena;
 
-    use super::*;
-    use crate::value::NumericalValue;
+    use crate::columnar::writer::column_operation::ColumnOperation;
+    use crate::{Cardinality, NumericalValue};
 
     #[test]
     fn test_column_writer_required_simple() {

columnar/src/columnar/writer/value_index.rs

@@ -45,16 +45,6 @@ impl<'a> SerializableOptionalIndex<'a> for SingleValueArrayIndex<'a> {
     }
 }
 
-impl OptionalIndexBuilder {
-    fn num_non_nulls(&self) -> u32 {
-        self.docs.len() as u32
-    }
-
-    fn iter(&self) -> Box<dyn Iterator<Item = u32> + '_> {
-        Box::new(self.docs.iter().copied())
-    }
-}
-
 impl OptionalIndexBuilder {
     pub fn finish<'a>(&'a mut self, num_rows: RowId) -> impl SerializableOptionalIndex + 'a {
         debug_assert!(self
@@ -96,7 +86,7 @@ pub struct MultivaluedIndexBuilder {
 impl MultivaluedIndexBuilder {
     pub fn finish(&mut self, num_docs: RowId) -> impl ColumnValues<u32> + '_ {
         self.start_offsets
-            .resize(num_docs as usize, self.total_num_vals_seen);
+            .resize(num_docs as usize + 1, self.total_num_vals_seen);
         VecColumn {
             values: &&self.start_offsets[..],
             min_value: 0,
@@ -188,7 +178,7 @@ mod tests {
                 .finish(4u32)
                 .iter()
                 .collect::<Vec<u32>>(),
-            vec![0, 0, 2, 3]
+            vec![0, 0, 2, 3, 3]
         );
         multivalued_value_index_builder.reset();
         multivalued_value_index_builder.record_row(2u32);
@@ -199,7 +189,7 @@ mod tests {
                 .finish(4u32)
                 .iter()
                 .collect::<Vec<u32>>(),
-            vec![0, 0, 0, 2]
+            vec![0, 0, 0, 2, 2]
         );
     }
 }

columnar/src/dynamic_column.rs

@@ -1,11 +1,14 @@
 use std::io;
-use std::net::IpAddr;
+use std::net::Ipv6Addr;
+use std::sync::Arc;
 
 use common::file_slice::FileSlice;
 use common::{HasLen, OwnedBytes};
 
-use crate::column::{BytesColumn, Column};
+use crate::column::{BytesColumn, Column, StrColumn};
+use crate::column_values::{monotonic_map_column, StrictlyMonotonicFn};
 use crate::columnar::ColumnType;
+use crate::{Cardinality, DateTime, NumericalType};
 
 #[derive(Clone)]
 pub enum DynamicColumn {
@@ -13,16 +16,151 @@ pub enum DynamicColumn {
     I64(Column<i64>),
     U64(Column<u64>),
     F64(Column<f64>),
-    IpAddr(Column<IpAddr>),
-    Str(BytesColumn),
-    DateTime(Column<crate::DateTime>),
+    IpAddr(Column<Ipv6Addr>),
+    DateTime(Column<DateTime>),
+    Bytes(BytesColumn),
+    Str(StrColumn),
+}
+
+impl DynamicColumn {
+    pub fn get_cardinality(&self) -> Cardinality {
+        match self {
+            DynamicColumn::Bool(c) => c.get_cardinality(),
+            DynamicColumn::I64(c) => c.get_cardinality(),
+            DynamicColumn::U64(c) => c.get_cardinality(),
+            DynamicColumn::F64(c) => c.get_cardinality(),
+            DynamicColumn::IpAddr(c) => c.get_cardinality(),
+            DynamicColumn::DateTime(c) => c.get_cardinality(),
+            DynamicColumn::Bytes(c) => c.ords().get_cardinality(),
+            DynamicColumn::Str(c) => c.ords().get_cardinality(),
+        }
+    }
+    pub fn column_type(&self) -> ColumnType {
+        match self {
+            DynamicColumn::Bool(_) => ColumnType::Bool,
+            DynamicColumn::I64(_) => ColumnType::I64,
+            DynamicColumn::U64(_) => ColumnType::U64,
+            DynamicColumn::F64(_) => ColumnType::F64,
+            DynamicColumn::IpAddr(_) => ColumnType::IpAddr,
+            DynamicColumn::DateTime(_) => ColumnType::DateTime,
+            DynamicColumn::Bytes(_) => ColumnType::Bytes,
+            DynamicColumn::Str(_) => ColumnType::Str,
+        }
+    }
+
+    pub fn is_numerical(&self) -> bool {
+        self.column_type().numerical_type().is_some()
+    }
+
+    pub fn is_f64(&self) -> bool {
+        self.column_type().numerical_type() == Some(NumericalType::F64)
+    }
+    pub fn is_i64(&self) -> bool {
+        self.column_type().numerical_type() == Some(NumericalType::I64)
+    }
+    pub fn is_u64(&self) -> bool {
+        self.column_type().numerical_type() == Some(NumericalType::U64)
+    }
+
+    pub fn coerce_to_f64(self) -> Option<DynamicColumn> {
+        match self {
+            DynamicColumn::I64(column) => Some(DynamicColumn::F64(Column {
+                idx: column.idx,
+                values: Arc::new(monotonic_map_column(column.values, MapI64ToF64)),
+            })),
+            DynamicColumn::U64(column) => Some(DynamicColumn::F64(Column {
+                idx: column.idx,
+                values: Arc::new(monotonic_map_column(column.values, MapU64ToF64)),
+            })),
+            DynamicColumn::F64(_) => Some(self),
+            _ => None,
+        }
+    }
+    pub fn coerce_to_i64(self) -> Option<DynamicColumn> {
+        match self {
+            DynamicColumn::U64(column) => {
+                if column.max_value() > i64::MAX as u64 {
+                    return None;
+                }
+                Some(DynamicColumn::I64(Column {
+                    idx: column.idx,
+                    values: Arc::new(monotonic_map_column(column.values, MapU64ToI64)),
+                }))
+            }
+            DynamicColumn::I64(_) => Some(self),
+            _ => None,
+        }
+    }
+    pub fn coerce_to_u64(self) -> Option<DynamicColumn> {
+        match self {
+            DynamicColumn::I64(column) => {
+                if column.min_value() < 0 {
+                    return None;
+                }
+                Some(DynamicColumn::U64(Column {
+                    idx: column.idx,
+                    values: Arc::new(monotonic_map_column(column.values, MapI64ToU64)),
+                }))
+            }
+            DynamicColumn::U64(_) => Some(self),
+            _ => None,
+        }
+    }
+}
+
+struct MapI64ToF64;
+impl StrictlyMonotonicFn<i64, f64> for MapI64ToF64 {
+    #[inline(always)]
+    fn mapping(&self, inp: i64) -> f64 {
+        inp as f64
+    }
+    #[inline(always)]
+    fn inverse(&self, out: f64) -> i64 {
+        out as i64
+    }
+}
+
+struct MapU64ToF64;
+impl StrictlyMonotonicFn<u64, f64> for MapU64ToF64 {
+    #[inline(always)]
+    fn mapping(&self, inp: u64) -> f64 {
+        inp as f64
+    }
+    #[inline(always)]
+    fn inverse(&self, out: f64) -> u64 {
+        out as u64
+    }
+}
+
+struct MapU64ToI64;
+impl StrictlyMonotonicFn<u64, i64> for MapU64ToI64 {
+    #[inline(always)]
+    fn mapping(&self, inp: u64) -> i64 {
+        inp as i64
+    }
+    #[inline(always)]
+    fn inverse(&self, out: i64) -> u64 {
+        out as u64
+    }
+}
+
+struct MapI64ToU64;
+impl StrictlyMonotonicFn<i64, u64> for MapI64ToU64 {
+    #[inline(always)]
+    fn mapping(&self, inp: i64) -> u64 {
+        inp as u64
+    }
+    #[inline(always)]
+    fn inverse(&self, out: u64) -> i64 {
+        out as i64
+    }
 }
 
 macro_rules! static_dynamic_conversions {
     ($typ:ty, $enum_name:ident) => {
-        impl Into<Option<Column<$typ>>> for DynamicColumn {
-            fn into(self) -> Option<Column<$typ>> {
-                if let Self::$enum_name(col) = self {
+        impl Into<Option<$typ>> for DynamicColumn {
+            fn into(self) -> Option<$typ> {
+                if let DynamicColumn::$enum_name(col) = self {
                     Some(col)
                 } else {
                     None
@@ -30,25 +168,22 @@ macro_rules! static_dynamic_conversions {
             }
         }
 
-        impl From<Column<$typ>> for DynamicColumn {
-            fn from(typed_column: Column<$typ>) -> Self {
+        impl From<$typ> for DynamicColumn {
+            fn from(typed_column: $typ) -> Self {
                 DynamicColumn::$enum_name(typed_column)
             }
         }
     };
 }
 
-static_dynamic_conversions!(bool, Bool);
-static_dynamic_conversions!(u64, U64);
-static_dynamic_conversions!(i64, I64);
-static_dynamic_conversions!(f64, F64);
-static_dynamic_conversions!(crate::DateTime, DateTime);
-
-impl From<BytesColumn> for DynamicColumn {
-    fn from(dictionary_encoded_col: BytesColumn) -> Self {
-        DynamicColumn::Str(dictionary_encoded_col)
-    }
-}
+static_dynamic_conversions!(Column<bool>, Bool);
+static_dynamic_conversions!(Column<u64>, U64);
+static_dynamic_conversions!(Column<i64>, I64);
+static_dynamic_conversions!(Column<f64>, F64);
+static_dynamic_conversions!(Column<crate::DateTime>, DateTime);
+static_dynamic_conversions!(StrColumn, Str);
+static_dynamic_conversions!(BytesColumn, Bytes);
+static_dynamic_conversions!(Column<Ipv6Addr>, IpAddr);
 
 #[derive(Clone)]
 pub struct DynamicColumnHandle {
@@ -77,12 +212,13 @@ impl DynamicColumnHandle {
     pub fn open_u64_lenient(&self) -> io::Result<Option<Column<u64>>> {
         let column_bytes = self.file_slice.read_bytes()?;
         match self.column_type {
-            ColumnType::Str => {
-                let column = crate::column::open_column_bytes(column_bytes)?;
+            ColumnType::Str | ColumnType::Bytes => {
+                let column: BytesColumn = crate::column::open_column_bytes(column_bytes)?;
                 Ok(Some(column.term_ord_column))
             }
             ColumnType::Bool => Ok(None),
-            ColumnType::Numerical(_) | ColumnType::DateTime => {
+            ColumnType::IpAddr => Ok(None),
+            ColumnType::I64 | ColumnType::U64 | ColumnType::F64 | ColumnType::DateTime => {
                 let column = crate::column::open_column_u64::<u64>(column_bytes)?;
                 Ok(Some(column))
             }
@@ -91,19 +227,15 @@ impl DynamicColumnHandle {
 
     fn open_internal(&self, column_bytes: OwnedBytes) -> io::Result<DynamicColumn> {
         let dynamic_column: DynamicColumn = match self.column_type {
-            ColumnType::Str => crate::column::open_column_bytes(column_bytes)?.into(),
-            ColumnType::Numerical(numerical_type) => match numerical_type {
-                crate::NumericalType::I64 => {
-                    crate::column::open_column_u64::<i64>(column_bytes)?.into()
-                }
-                crate::NumericalType::U64 => {
-                    crate::column::open_column_u64::<u64>(column_bytes)?.into()
-                }
-                crate::NumericalType::F64 => {
-                    crate::column::open_column_u64::<f64>(column_bytes)?.into()
-                }
-            },
+            ColumnType::Bytes => {
+                crate::column::open_column_bytes::<BytesColumn>(column_bytes)?.into()
+            }
+            ColumnType::Str => crate::column::open_column_bytes::<StrColumn>(column_bytes)?.into(),
+            ColumnType::I64 => crate::column::open_column_u64::<i64>(column_bytes)?.into(),
+            ColumnType::U64 => crate::column::open_column_u64::<u64>(column_bytes)?.into(),
+            ColumnType::F64 => crate::column::open_column_u64::<f64>(column_bytes)?.into(),
             ColumnType::Bool => crate::column::open_column_u64::<bool>(column_bytes)?.into(),
+            ColumnType::IpAddr => crate::column::open_column_u128::<Ipv6Addr>(column_bytes)?.into(),
             ColumnType::DateTime => {
                 crate::column::open_column_u64::<crate::DateTime>(column_bytes)?.into()
             }

columnar/src/lib.rs

@@ -18,9 +18,12 @@ mod dynamic_column;
 pub(crate) mod utils;
 mod value;
 
-pub use column::Column;
+pub use column::{BytesColumn, Column, StrColumn};
 pub use column_values::ColumnValues;
-pub use columnar::{ColumnType, ColumnarReader, ColumnarWriter, HasAssociatedColumnType};
+pub use columnar::{
+    merge_columnar, ColumnType, ColumnarReader, ColumnarWriter, HasAssociatedColumnType,
+    MergeDocOrder,
+};
 pub use value::{NumericalType, NumericalValue};
 
 pub use self::dynamic_column::{DynamicColumn, DynamicColumnHandle};
@@ -59,6 +62,12 @@ pub enum Cardinality {
 }
 
 impl Cardinality {
+    pub fn is_optional(&self) -> bool {
+        matches!(self, Cardinality::Optional)
+    }
+    pub fn is_multivalue(&self) -> bool {
+        matches!(self, Cardinality::Multivalued)
+    }
     pub(crate) fn to_code(self) -> u8 {
         self as u8
     }

columnar/src/tests.rs

@@ -1,10 +1,13 @@
+use std::net::Ipv6Addr;
+
+use crate::column_values::MonotonicallyMappableToU128;
 use crate::columnar::ColumnType;
 use crate::dynamic_column::{DynamicColumn, DynamicColumnHandle};
 use crate::value::NumericalValue;
 use crate::{Cardinality, ColumnarReader, ColumnarWriter};
 
 #[test]
-fn test_dataframe_writer_bytes() {
+fn test_dataframe_writer_str() {
     let mut dataframe_writer = ColumnarWriter::default();
     dataframe_writer.record_str(1u32, "my_string", "hello");
     dataframe_writer.record_str(3u32, "my_string", "helloeee");
@@ -14,7 +17,21 @@ fn test_dataframe_writer_bytes() {
     assert_eq!(columnar.num_columns(), 1);
     let cols: Vec<DynamicColumnHandle> = columnar.read_columns("my_string").unwrap();
     assert_eq!(cols.len(), 1);
-    assert_eq!(cols[0].num_bytes(), 165);
+    assert_eq!(cols[0].num_bytes(), 158);
+}
+
+#[test]
+fn test_dataframe_writer_bytes() {
+    let mut dataframe_writer = ColumnarWriter::default();
+    dataframe_writer.record_bytes(1u32, "my_string", b"hello");
+    dataframe_writer.record_bytes(3u32, "my_string", b"helloeee");
+    let mut buffer: Vec<u8> = Vec::new();
+    dataframe_writer.serialize(5, &mut buffer).unwrap();
+    let columnar = ColumnarReader::open(buffer).unwrap();
+    assert_eq!(columnar.num_columns(), 1);
+    let cols: Vec<DynamicColumnHandle> = columnar.read_columns("my_string").unwrap();
+    assert_eq!(cols.len(), 1);
+    assert_eq!(cols[0].num_bytes(), 158);
 }
 
 #[test]
@@ -28,7 +45,7 @@ fn test_dataframe_writer_bool() {
     assert_eq!(columnar.num_columns(), 1);
     let cols: Vec<DynamicColumnHandle> = columnar.read_columns("bool.value").unwrap();
     assert_eq!(cols.len(), 1);
-    assert_eq!(cols[0].num_bytes(), 29);
+    assert_eq!(cols[0].num_bytes(), 22);
     assert_eq!(cols[0].column_type(), ColumnType::Bool);
     let dyn_bool_col = cols[0].open().unwrap();
     let DynamicColumn::Bool(bool_col) = dyn_bool_col else { panic!(); };
@@ -36,6 +53,59 @@ fn test_dataframe_writer_bool() {
     assert_eq!(&vals, &[None, Some(false), None, Some(true), None,]);
 }
 
+#[test]
+fn test_dataframe_writer_u64_multivalued() {
+    let mut dataframe_writer = ColumnarWriter::default();
+    dataframe_writer.record_numerical(2u32, "divisor", 2u64);
+    dataframe_writer.record_numerical(3u32, "divisor", 3u64);
+    dataframe_writer.record_numerical(4u32, "divisor", 2u64);
+    dataframe_writer.record_numerical(5u32, "divisor", 5u64);
+    dataframe_writer.record_numerical(6u32, "divisor", 2u64);
+    dataframe_writer.record_numerical(6u32, "divisor", 3u64);
+    let mut buffer: Vec<u8> = Vec::new();
+    dataframe_writer.serialize(7, &mut buffer).unwrap();
+    let columnar = ColumnarReader::open(buffer).unwrap();
+    assert_eq!(columnar.num_columns(), 1);
+    let cols: Vec<DynamicColumnHandle> = columnar.read_columns("divisor").unwrap();
+    assert_eq!(cols.len(), 1);
+    assert_eq!(cols[0].num_bytes(), 29);
+    let dyn_i64_col = cols[0].open().unwrap();
+    let DynamicColumn::I64(divisor_col) = dyn_i64_col else { panic!(); };
+    assert_eq!(
+        divisor_col.get_cardinality(),
+        crate::Cardinality::Multivalued
+    );
+    assert_eq!(divisor_col.num_rows(), 7);
+}
+
+#[test]
+fn test_dataframe_writer_ip_addr() {
+    let mut dataframe_writer = ColumnarWriter::default();
+    dataframe_writer.record_ip_addr(1, "ip_addr", Ipv6Addr::from_u128(1001));
+    dataframe_writer.record_ip_addr(3, "ip_addr", Ipv6Addr::from_u128(1050));
+    let mut buffer: Vec<u8> = Vec::new();
+    dataframe_writer.serialize(5, &mut buffer).unwrap();
+    let columnar = ColumnarReader::open(buffer).unwrap();
+    assert_eq!(columnar.num_columns(), 1);
+    let cols: Vec<DynamicColumnHandle> = columnar.read_columns("ip_addr").unwrap();
+    assert_eq!(cols.len(), 1);
+    assert_eq!(cols[0].num_bytes(), 42);
+    assert_eq!(cols[0].column_type(), ColumnType::IpAddr);
+    let dyn_bool_col = cols[0].open().unwrap();
+    let DynamicColumn::IpAddr(ip_col) = dyn_bool_col else { panic!(); };
+    let vals: Vec<Option<Ipv6Addr>> = (0..5).map(|row_id| ip_col.first(row_id)).collect();
+    assert_eq!(
+        &vals,
+        &[
+            None,
+            Some(Ipv6Addr::from_u128(1001)),
+            None,
+            Some(Ipv6Addr::from_u128(1050)),
+            None,
+        ]
+    );
+}
+
 #[test]
 fn test_dataframe_writer_numerical() {
     let mut dataframe_writer = ColumnarWriter::default();
@@ -53,7 +123,7 @@ fn test_dataframe_writer_numerical() {
     // - header 14 bytes
     // - vals  8 //< due to padding? could have been 1byte?.
     // - null footer 6 bytes
-    assert_eq!(cols[0].num_bytes(), 40);
+    assert_eq!(cols[0].num_bytes(), 33);
     let column = cols[0].open().unwrap();
     let DynamicColumn::I64(column_i64) = column else { panic!(); };
     assert_eq!(column_i64.idx.get_cardinality(), Cardinality::Optional);
@@ -67,18 +137,76 @@ fn test_dataframe_writer_numerical() {
 }
 
 #[test]
-fn test_dictionary_encoded() {
+fn test_dictionary_encoded_str() {
     let mut buffer = Vec::new();
     let mut columnar_writer = ColumnarWriter::default();
-    columnar_writer.record_str(1, "my.column", "my.key");
-    columnar_writer.record_str(3, "my.column", "my.key2");
+    columnar_writer.record_str(1, "my.column", "a");
+    columnar_writer.record_str(3, "my.column", "c");
     columnar_writer.record_str(3, "my.column2", "different_column!");
+    columnar_writer.record_str(4, "my.column", "b");
     columnar_writer.serialize(5, &mut buffer).unwrap();
     let columnar_reader = ColumnarReader::open(buffer).unwrap();
     assert_eq!(columnar_reader.num_columns(), 2);
     let col_handles = columnar_reader.read_columns("my.column").unwrap();
     assert_eq!(col_handles.len(), 1);
     let DynamicColumn::Str(str_col) = col_handles[0].open().unwrap() else  { panic!(); };
+    let index: Vec<Option<u64>> = (0..5).map(|row_id| str_col.ords().first(row_id)).collect();
+    assert_eq!(index, &[None, Some(0), None, Some(2), Some(1)]);
     assert_eq!(str_col.num_rows(), 5);
-    // let term_ords = (0..)
+    let mut term_buffer = String::new();
+    let term_ords = str_col.ords();
+    assert_eq!(term_ords.first(0), None);
+    assert_eq!(term_ords.first(1), Some(0));
+    str_col.ord_to_str(0u64, &mut term_buffer).unwrap();
+    assert_eq!(term_buffer, "a");
+    assert_eq!(term_ords.first(2), None);
+    assert_eq!(term_ords.first(3), Some(2));
+    str_col.ord_to_str(2u64, &mut term_buffer).unwrap();
+    assert_eq!(term_buffer, "c");
+    assert_eq!(term_ords.first(4), Some(1));
+    str_col.ord_to_str(1u64, &mut term_buffer).unwrap();
+    assert_eq!(term_buffer, "b");
+}
+
+#[test]
+fn test_dictionary_encoded_bytes() {
+    let mut buffer = Vec::new();
+    let mut columnar_writer = ColumnarWriter::default();
+    columnar_writer.record_bytes(1, "my.column", b"a");
+    columnar_writer.record_bytes(3, "my.column", b"c");
+    columnar_writer.record_bytes(3, "my.column2", b"different_column!");
+    columnar_writer.record_bytes(4, "my.column", b"b");
+    columnar_writer.serialize(5, &mut buffer).unwrap();
+    let columnar_reader = ColumnarReader::open(buffer).unwrap();
+    assert_eq!(columnar_reader.num_columns(), 2);
+    let col_handles = columnar_reader.read_columns("my.column").unwrap();
+    assert_eq!(col_handles.len(), 1);
+    let DynamicColumn::Bytes(bytes_col) = col_handles[0].open().unwrap() else  { panic!(); };
+    let index: Vec<Option<u64>> = (0..5)
+        .map(|row_id| bytes_col.ords().first(row_id))
+        .collect();
+    assert_eq!(index, &[None, Some(0), None, Some(2), Some(1)]);
+    assert_eq!(bytes_col.num_rows(), 5);
+    let mut term_buffer = Vec::new();
+    let term_ords = bytes_col.ords();
+    assert_eq!(term_ords.first(0), None);
+    assert_eq!(term_ords.first(1), Some(0));
+    bytes_col
+        .dictionary
+        .ord_to_term(0u64, &mut term_buffer)
+        .unwrap();
+    assert_eq!(term_buffer, b"a");
+    assert_eq!(term_ords.first(2), None);
+    assert_eq!(term_ords.first(3), Some(2));
+    bytes_col
+        .dictionary
+        .ord_to_term(2u64, &mut term_buffer)
+        .unwrap();
+    assert_eq!(term_buffer, b"c");
+    assert_eq!(term_ords.first(4), Some(1));
+    bytes_col
+        .dictionary
+        .ord_to_term(1u64, &mut term_buffer)
+        .unwrap();
+    assert_eq!(term_buffer, b"b");
 }

columnar/src/value.rs

@@ -1,12 +1,22 @@
-use crate::{Column, ColumnType, InvalidData};
+use crate::InvalidData;
 
-#[derive(Copy, Clone, Debug, PartialEq)]
+#[derive(Copy, Clone, PartialEq, Debug)]
 pub enum NumericalValue {
     I64(i64),
     U64(u64),
     F64(f64),
 }
 
+impl NumericalValue {
+    pub fn numerical_type(&self) -> NumericalType {
+        match self {
+            NumericalValue::I64(_) => NumericalType::I64,
+            NumericalValue::U64(_) => NumericalType::U64,
+            NumericalValue::F64(_) => NumericalType::F64,
+        }
+    }
+}
+
 impl From<u64> for NumericalValue {
     fn from(val: u64) -> NumericalValue {
         NumericalValue::U64(val)
@@ -25,18 +35,6 @@ impl From<f64> for NumericalValue {
     }
 }
 
-impl NumericalValue {
-    pub fn numerical_type(&self) -> NumericalType {
-        match self {
-            NumericalValue::F64(_) => NumericalType::F64,
-            NumericalValue::I64(_) => NumericalType::I64,
-            NumericalValue::U64(_) => NumericalType::U64,
-        }
-    }
-}
-
-impl Eq for NumericalValue {}
-
 #[derive(Clone, Copy, Debug, Default, Hash, Eq, PartialEq)]
 #[repr(u8)]
 pub enum NumericalType {

examples/aggregation.rs

@@ -13,7 +13,7 @@ use tantivy::aggregation::agg_result::AggregationResults;
 use tantivy::aggregation::metric::AverageAggregation;
 use tantivy::aggregation::AggregationCollector;
 use tantivy::query::TermQuery;
-use tantivy::schema::{self, IndexRecordOption, Schema, TextFieldIndexing};
+use tantivy::schema::{self, Cardinality, IndexRecordOption, Schema, TextFieldIndexing};
 use tantivy::{doc, Index, Term};
 
 fn main() -> tantivy::Result<()> {
@@ -25,7 +25,7 @@ fn main() -> tantivy::Result<()> {
         .set_stored();
     let text_field = schema_builder.add_text_field("text", text_fieldtype);
     let score_fieldtype =
-        crate::schema::NumericOptions::default().set_fast();
+        crate::schema::NumericOptions::default().set_fast(Cardinality::SingleValue);
     let highscore_field = schema_builder.add_f64_field("highscore", score_fieldtype.clone());
     let price_field = schema_builder.add_f64_field("price", score_fieldtype);
 

examples/date_time_field.rs

@@ -4,7 +4,7 @@
 
 use tantivy::collector::TopDocs;
 use tantivy::query::QueryParser;
-use tantivy::schema::{DateOptions, Schema, Value, INDEXED, STORED, STRING};
+use tantivy::schema::{Cardinality, DateOptions, Schema, Value, INDEXED, STORED, STRING};
 use tantivy::Index;
 
 fn main() -> tantivy::Result<()> {
@@ -12,7 +12,7 @@ fn main() -> tantivy::Result<()> {
     let mut schema_builder = Schema::builder();
     let opts = DateOptions::from(INDEXED)
         .set_stored()
-        .set_fast()
+        .set_fast(Cardinality::SingleValue)
         .set_precision(tantivy::DatePrecision::Seconds);
     let occurred_at = schema_builder.add_date_field("occurred_at", opts);
     let event_type = schema_builder.add_text_field("event", STRING | STORED);

fastfield_codecs/Cargo.toml

@@ -14,7 +14,6 @@ repository = "https://github.com/quickwit-oss/tantivy"
 [dependencies]
 common = { version = "0.5", path = "../common/", package = "tantivy-common" }
 tantivy-bitpacker = { version= "0.3", path = "../bitpacker/" }
-columnar = { version= "0.1", path="../columnar", package="tantivy-columnar" }
 prettytable-rs = {version="0.10.0", optional= true}
 rand = {version="0.8.3", optional= true}
 fastdivide = "0.4"

fastfield_codecs/src/column.rs

@@ -2,11 +2,81 @@ use std::fmt::{self, Debug};
 use std::marker::PhantomData;
 use std::ops::{Range, RangeInclusive};
 
-pub use columnar::ColumnValues as Column;
 use tantivy_bitpacker::minmax;
 
 use crate::monotonic_mapping::StrictlyMonotonicFn;
 
+/// `Column` provides columnar access on a field.
+pub trait Column<T: PartialOrd + Debug = u64>: Send + Sync {
+    /// Return the value associated with the given idx.
+    ///
+    /// This accessor should return as fast as possible.
+    ///
+    /// # Panics
+    ///
+    /// May panic if `idx` is greater than the column length.
+    fn get_val(&self, idx: u32) -> T;
+
+    /// Fills an output buffer with the fast field values
+    /// associated with the `DocId` going from
+    /// `start` to `start + output.len()`.
+    ///
+    /// # Panics
+    ///
+    /// Must panic if `start + output.len()` is greater than
+    /// the segment's `maxdoc`.
+    #[inline]
+    fn get_range(&self, start: u64, output: &mut [T]) {
+        for (out, idx) in output.iter_mut().zip(start..) {
+            *out = self.get_val(idx as u32);
+        }
+    }
+
+    /// Get the positions of values which are in the provided value range.
+    ///
+    /// Note that position == docid for single value fast fields
+    #[inline]
+    fn get_docids_for_value_range(
+        &self,
+        value_range: RangeInclusive<T>,
+        doc_id_range: Range<u32>,
+        positions: &mut Vec<u32>,
+    ) {
+        let doc_id_range = doc_id_range.start..doc_id_range.end.min(self.num_vals());
+
+        for idx in doc_id_range.start..doc_id_range.end {
+            let val = self.get_val(idx);
+            if value_range.contains(&val) {
+                positions.push(idx);
+            }
+        }
+    }
+
+    /// Returns the minimum value for this fast field.
+    ///
+    /// This min_value may not be exact.
+    /// For instance, the min value does not take in account of possible
+    /// deleted document. All values are however guaranteed to be higher than
+    /// `.min_value()`.
+    fn min_value(&self) -> T;
+
+    /// Returns the maximum value for this fast field.
+    ///
+    /// This max_value may not be exact.
+    /// For instance, the max value does not take in account of possible
+    /// deleted document. All values are however guaranteed to be higher than
+    /// `.max_value()`.
+    fn max_value(&self) -> T;
+
+    /// The number of values in the column.
+    fn num_vals(&self) -> u32;
+
+    /// Returns a iterator over the data
+    fn iter<'a>(&'a self) -> Box<dyn Iterator<Item = T> + 'a> {
+        Box::new((0..self.num_vals()).map(|idx| self.get_val(idx)))
+    }
+}
+
 /// VecColumn provides `Column` over a slice.
 pub struct VecColumn<'a, T = u64> {
     values: &'a [T],
@@ -14,6 +84,32 @@ pub struct VecColumn<'a, T = u64> {
     max_value: T,
 }
 
+impl<'a, C: Column<T>, T: Copy + PartialOrd + fmt::Debug> Column<T> for &'a C {
+    fn get_val(&self, idx: u32) -> T {
+        (*self).get_val(idx)
+    }
+
+    fn min_value(&self) -> T {
+        (*self).min_value()
+    }
+
+    fn max_value(&self) -> T {
+        (*self).max_value()
+    }
+
+    fn num_vals(&self) -> u32 {
+        (*self).num_vals()
+    }
+
+    fn iter<'b>(&'b self) -> Box<dyn Iterator<Item = T> + 'b> {
+        (*self).iter()
+    }
+
+    fn get_range(&self, start: u64, output: &mut [T]) {
+        (*self).get_range(start, output)
+    }
+}
+
 impl<'a, T: Copy + PartialOrd + Send + Sync + Debug> Column<T> for VecColumn<'a, T> {
     fn get_val(&self, position: u32) -> T {
         self.values[position as usize]

fastfield_codecs/src/serialize.rs

@@ -402,8 +402,8 @@ mod tests {
         let mut buffer = Vec::new();
         let col = VecColumn::from(&[false, true][..]);
         serialize(col, &mut buffer, &ALL_CODEC_TYPES).unwrap();
-        // 5 bytes of header, 1 byte of value, 7 bytes of padding.
-        assert_eq!(buffer.len(), 3 + 5 + 8 + 4 + 2);
+        // 5 bytes of header, 1 byte of value
+        assert_eq!(buffer.len(), 3 + 5 + 1 + 4 + 2);
     }
 
     #[test]
@@ -411,8 +411,8 @@ mod tests {
         let mut buffer = Vec::new();
         let col = VecColumn::from(&[true][..]);
         serialize(col, &mut buffer, &ALL_CODEC_TYPES).unwrap();
-        // 5 bytes of header, 0 bytes of value, 7 bytes of padding.
-        assert_eq!(buffer.len(), 3 + 5 + 7 + 4 + 2);
+        // 5 bytes of header, 0 bytes of value
+        assert_eq!(buffer.len(), 3 + 5 + 4 + 2);
     }
 
     #[test]
@@ -422,6 +422,6 @@ mod tests {
         let col = VecColumn::from(&vals[..]);
         serialize(col, &mut buffer, &[FastFieldCodecType::Bitpacked]).unwrap();
         // Values are stored over 3 bits.
-        assert_eq!(buffer.len(), 3 + 7 + (3 * 80 / 8) + 7 + 4 + 2);
+        assert_eq!(buffer.len(), 3 + 7 + (3 * 80 / 8) + 4 + 2);
     }
 }

run-tests.sh

@@ -1,2 +0,0 @@
-#!/bin/bash
-cargo test

src/aggregation/agg_req_with_accessor.rs

@@ -15,7 +15,7 @@ use super::metric::{
 use super::segment_agg_result::BucketCount;
 use super::VecWithNames;
 use crate::fastfield::{type_and_cardinality, MultiValuedFastFieldReader};
-use crate::schema::Type;
+use crate::schema::{Cardinality, Type};
 use crate::{InvertedIndexReader, SegmentReader, TantivyError};
 
 #[derive(Clone, Default)]

src/aggregation/metric/average.rs

@@ -6,14 +6,13 @@ use super::{IntermediateStats, SegmentStatsCollector};
 
 /// A single-value metric aggregation that computes the average of numeric values that are
 /// extracted from the aggregated documents.
-/// Supported field types are u64, i64, and f64.
 /// See [super::SingleMetricResult] for return value.
 ///
 /// # JSON Format
 /// ```json
 /// {
 ///     "avg": {
-///         "field": "score",
+///         "field": "score"
 ///     }
 /// }
 /// ```

src/aggregation/metric/count.rs

@@ -6,14 +6,13 @@ use super::{IntermediateStats, SegmentStatsCollector};
 
 /// A single-value metric aggregation that counts the number of values that are
 /// extracted from the aggregated documents.
-/// Supported field types are u64, i64, and f64.
 /// See [super::SingleMetricResult] for return value.
 ///
 /// # JSON Format
 /// ```json
 /// {
 ///     "value_count": {
-///         "field": "score",
+///         "field": "score"
 ///     }
 /// }
 /// ```

src/aggregation/metric/max.rs

@@ -6,14 +6,13 @@ use super::{IntermediateStats, SegmentStatsCollector};
 
 /// A single-value metric aggregation that computes the maximum of numeric values that are
 /// extracted from the aggregated documents.
-/// Supported field types are u64, i64, and f64.
 /// See [super::SingleMetricResult] for return value.
 ///
 /// # JSON Format
 /// ```json
 /// {
 ///     "max": {
-///         "field": "score",
+///         "field": "score"
 ///     }
 /// }
 /// ```

src/aggregation/metric/min.rs

@@ -6,14 +6,13 @@ use super::{IntermediateStats, SegmentStatsCollector};
 
 /// A single-value metric aggregation that computes the minimum of numeric values that are
 /// extracted from the aggregated documents.
-/// Supported field types are u64, i64, and f64.
 /// See [super::SingleMetricResult] for return value.
 ///
 /// # JSON Format
 /// ```json
 /// {
 ///     "min": {
-///         "field": "score",
+///         "field": "score"
 ///     }
 /// }
 /// ```

src/aggregation/metric/mod.rs

@@ -43,13 +43,13 @@ mod tests {
     use crate::aggregation::agg_result::AggregationResults;
     use crate::aggregation::AggregationCollector;
     use crate::query::AllQuery;
-    use crate::schema::{NumericOptions, Schema};
+    use crate::schema::{Cardinality, NumericOptions, Schema};
     use crate::Index;
 
     #[test]
     fn test_metric_aggregations() {
         let mut schema_builder = Schema::builder();
-        let field_options = NumericOptions::default().set_fast();
+        let field_options = NumericOptions::default().set_fast(Cardinality::SingleValue);
         let field = schema_builder.add_f64_field("price", field_options);
         let index = Index::create_in_ram(schema_builder.build());
         let mut index_writer = index.writer_for_tests().unwrap();

src/aggregation/metric/stats.rs

@@ -7,14 +7,13 @@ use crate::{DocId, TantivyError};
 
 /// A multi-value metric aggregation that computes a collection of statistics on numeric values that
 /// are extracted from the aggregated documents.
-/// Supported field types are `u64`, `i64`, and `f64`.
 /// See [`Stats`] for returned statistics.
 ///
 /// # JSON Format
 /// ```json
 /// {
 ///     "stats": {
-///         "field": "score",
+///         "field": "score"
 ///     }
 ///  }
 /// ```

src/aggregation/metric/sum.rs

@@ -6,14 +6,13 @@ use super::{IntermediateStats, SegmentStatsCollector};
 
 /// A single-value metric aggregation that sums up numeric values that are
 /// extracted from the aggregated documents.
-/// Supported field types are u64, i64, and f64.
 /// See [super::SingleMetricResult] for return value.
 ///
 /// # JSON Format
 /// ```json
 /// {
 ///     "sum": {
-///         "field": "score",
+///         "field": "score"
 ///     }
 /// }
 /// ```

src/aggregation/mod.rs

@@ -1,6 +1,5 @@
 //! # Aggregations
 //!
-//!
 //! An aggregation summarizes your data as statistics on buckets or metrics.
 //!
 //! Aggregations can provide answer to questions like:
@@ -41,6 +40,10 @@
 //! - [Metric](metric)
 //!     - [Average](metric::AverageAggregation)
 //!     - [Stats](metric::StatsAggregation)
+//!     - [Min](metric::MinAggregation)
+//!     - [Max](metric::MaxAggregation)
+//!     - [Sum](metric::SumAggregation)
+//!     - [Count](metric::CountAggregation)
 //!
 //! # Example
 //! Compute the average metric, by building [`agg_req::Aggregations`], which is built from an
@@ -75,7 +78,7 @@
 //! }
 //! ```
 //! # Example JSON
-//! Requests are compatible with the elasticsearch json request format.
+//! Requests are compatible with the elasticsearch JSON request format.
 //!
 //! ```
 //! use tantivy::aggregation::agg_req::Aggregations;
@@ -430,13 +433,13 @@ mod tests {
         let text_field_id = schema_builder.add_text_field("text_id", text_fieldtype);
         let string_field_id = schema_builder.add_text_field("string_id", STRING | FAST);
         let score_fieldtype =
-            crate::schema::NumericOptions::default().set_fast();
+            crate::schema::NumericOptions::default().set_fast(Cardinality::SingleValue);
         let score_field = schema_builder.add_u64_field("score", score_fieldtype.clone());
         let score_field_f64 = schema_builder.add_f64_field("score_f64", score_fieldtype.clone());
         let score_field_i64 = schema_builder.add_i64_field("score_i64", score_fieldtype);
         let fraction_field = schema_builder.add_f64_field(
             "fraction_f64",
-            crate::schema::NumericOptions::default().set_fast(),
+            crate::schema::NumericOptions::default().set_fast(Cardinality::SingleValue),
         );
         let index = Index::create_in_ram(schema_builder.build());
         {
@@ -654,12 +657,12 @@ mod tests {
         let date_field = schema_builder.add_date_field("date", FAST);
         schema_builder.add_text_field("dummy_text", STRING);
         let score_fieldtype =
-            crate::schema::NumericOptions::default().set_fast();
+            crate::schema::NumericOptions::default().set_fast(Cardinality::SingleValue);
         let score_field = schema_builder.add_u64_field("score", score_fieldtype.clone());
         let score_field_f64 = schema_builder.add_f64_field("score_f64", score_fieldtype.clone());
 
         let multivalue =
-            crate::schema::NumericOptions::default().set_fast();
+            crate::schema::NumericOptions::default().set_fast(Cardinality::MultiValues);
         let scores_field_i64 = schema_builder.add_i64_field("scores_i64", multivalue);
 
         let score_field_i64 = schema_builder.add_i64_field("score_i64", score_fieldtype);
@@ -1187,7 +1190,7 @@ mod tests {
             let text_field_few_terms =
                 schema_builder.add_text_field("text_few_terms", STRING | FAST);
             let score_fieldtype =
-                crate::schema::NumericOptions::default().set_fast();
+                crate::schema::NumericOptions::default().set_fast(Cardinality::SingleValue);
             let score_field = schema_builder.add_u64_field("score", score_fieldtype.clone());
             let score_field_f64 =
                 schema_builder.add_f64_field("score_f64", score_fieldtype.clone());

src/collector/filter_collector_wrapper.rs

@@ -12,10 +12,10 @@
 use std::marker::PhantomData;
 use std::sync::Arc;
 
-use columnar::{DynamicColumn, HasAssociatedColumnType};
 use fastfield_codecs::Column;
 
 use crate::collector::{Collector, SegmentCollector};
+use crate::fastfield::FastValue;
 use crate::schema::Field;
 use crate::{Score, SegmentReader, TantivyError};
 
@@ -61,7 +61,7 @@ use crate::{Score, SegmentReader, TantivyError};
 /// # Ok(())
 /// # }
 /// ```
-pub struct FilterCollector<TCollector, TPredicate, TPredicateValue: Default>
+pub struct FilterCollector<TCollector, TPredicate, TPredicateValue: FastValue>
 where TPredicate: 'static + Clone
 {
     field: Field,
@@ -70,7 +70,7 @@ where TPredicate: 'static + Clone
     t_predicate_value: PhantomData<TPredicateValue>,
 }
 
-impl<TCollector, TPredicate, TPredicateValue: Default>
+impl<TCollector, TPredicate, TPredicateValue: FastValue>
     FilterCollector<TCollector, TPredicate, TPredicateValue>
 where
     TCollector: Collector + Send + Sync,
@@ -91,13 +91,12 @@ where
     }
 }
 
-impl<TCollector, TPredicate, TPredicateValue: Default> Collector
+impl<TCollector, TPredicate, TPredicateValue: FastValue> Collector
     for FilterCollector<TCollector, TPredicate, TPredicateValue>
 where
     TCollector: Collector + Send + Sync,
     TPredicate: 'static + Fn(TPredicateValue) -> bool + Send + Sync + Clone,
-    TPredicateValue: HasAssociatedColumnType,
-    DynamicColumn: Into<Option<columnar::Column<TPredicateValue>>>,
+    TPredicateValue: FastValue,
 {
     // That's the type of our result.
     // Our standard deviation will be a float.
@@ -118,10 +117,20 @@ where
                 field_entry.name()
             )));
         }
+        let requested_type = TPredicateValue::to_type();
+        let field_schema_type = field_entry.field_type().value_type();
+        if requested_type != field_schema_type {
+            return Err(TantivyError::SchemaError(format!(
+                "Field {:?} is of type {:?}!={:?}",
+                field_entry.name(),
+                requested_type,
+                field_schema_type
+            )));
+        }
 
         let fast_field_reader = segment_reader
             .fast_fields()
-            .typed_column_first_or_default(schema.get_field_name(self.field))?;
+            .typed_fast_field_reader(schema.get_field_name(self.field))?;
 
         let segment_collector = self
             .collector
@@ -150,7 +159,7 @@ where
 pub struct FilterSegmentCollector<TSegmentCollector, TPredicate, TPredicateValue>
 where
     TPredicate: 'static,
-    DynamicColumn: Into<Option<columnar::Column<TPredicateValue>>>,
+    TPredicateValue: FastValue,
 {
     fast_field_reader: Arc<dyn Column<TPredicateValue>>,
     segment_collector: TSegmentCollector,
@@ -162,9 +171,8 @@ impl<TSegmentCollector, TPredicate, TPredicateValue> SegmentCollector
     for FilterSegmentCollector<TSegmentCollector, TPredicate, TPredicateValue>
 where
     TSegmentCollector: SegmentCollector,
-    TPredicateValue: HasAssociatedColumnType,
     TPredicate: 'static + Fn(TPredicateValue) -> bool + Send + Sync,
-    DynamicColumn: Into<Option<columnar::Column<TPredicateValue>>>,
+    TPredicateValue: FastValue,
 {
     type Fruit = TSegmentCollector::Fruit;
 

src/collector/histogram_collector.rs

@@ -4,7 +4,7 @@ use fastdivide::DividerU64;
 use fastfield_codecs::Column;
 
 use crate::collector::{Collector, SegmentCollector};
-use crate::fastfield::{FastFieldNotAvailableError, FastValue};
+use crate::fastfield::FastValue;
 use crate::schema::Type;
 use crate::{DocId, Score};
 
@@ -87,14 +87,14 @@ impl HistogramComputer {
 }
 pub struct SegmentHistogramCollector {
     histogram_computer: HistogramComputer,
-    column_u64: Arc<dyn Column<u64>>,
+    ff_reader: Arc<dyn Column<u64>>,
 }
 
 impl SegmentCollector for SegmentHistogramCollector {
     type Fruit = Vec<u64>;
 
     fn collect(&mut self, doc: DocId, _score: Score) {
-        let value = self.column_u64.get_val(doc);
+        let value = self.ff_reader.get_val(doc);
         self.histogram_computer.add_value(value);
     }
 
@@ -112,18 +112,14 @@ impl Collector for HistogramCollector {
         _segment_local_id: crate::SegmentOrdinal,
         segment: &crate::SegmentReader,
     ) -> crate::Result<Self::Child> {
-        let column_opt = segment.fast_fields().u64_lenient(&self.field)?;
-        let column = column_opt.ok_or_else(|| FastFieldNotAvailableError {
-            field_name: self.field.clone(),
-        })?;
-        let column_u64 = column.first_or_default_col(0u64);
+        let ff_reader = segment.fast_fields().u64_lenient(&self.field)?;
         Ok(SegmentHistogramCollector {
             histogram_computer: HistogramComputer {
                 counts: vec![0; self.num_buckets],
                 min_value: self.min_value,
                 divider: self.divider,
             },
-            column_u64,
+            ff_reader,
         })
     }
 

src/collector/mod.rs

@@ -104,8 +104,9 @@ pub use self::custom_score_top_collector::{CustomScorer, CustomSegmentScorer};
 
 mod tweak_score_top_collector;
 pub use self::tweak_score_top_collector::{ScoreSegmentTweaker, ScoreTweaker};
-// mod facet_collector;
-// pub use self::facet_collector::{FacetCollector, FacetCounts};
+
+mod facet_collector;
+pub use self::facet_collector::{FacetCollector, FacetCounts};
 use crate::query::Weight;
 
 mod docset_collector;

src/collector/tests.rs

@@ -5,6 +5,7 @@ use fastfield_codecs::Column;
 use super::*;
 use crate::collector::{Count, FilterCollector, TopDocs};
 use crate::core::SegmentReader;
+use crate::fastfield::BytesFastFieldReader;
 use crate::query::{AllQuery, QueryParser};
 use crate::schema::{Field, Schema, FAST, TEXT};
 use crate::time::format_description::well_known::Rfc3339;
@@ -57,10 +58,9 @@ pub fn test_filter_collector() -> crate::Result<()> {
 
     assert_eq!(filtered_top_docs.len(), 0);
 
-    fn date_filter(value: columnar::DateTime) -> bool {
-        (crate::DateTime::from(value).into_utc()
-            - OffsetDateTime::parse("2019-04-09T00:00:00+00:00", &Rfc3339).unwrap())
-        .whole_weeks()
+    fn date_filter(value: DateTime) -> bool {
+        (value.into_utc() - OffsetDateTime::parse("2019-04-09T00:00:00+00:00", &Rfc3339).unwrap())
+            .whole_weeks()
             > 0
     }
 
@@ -164,10 +164,8 @@ pub struct FastFieldSegmentCollector {
 }
 
 impl FastFieldTestCollector {
-    pub fn for_field(field: impl ToString) -> FastFieldTestCollector {
-        FastFieldTestCollector {
-            field: field.to_string(),
-        }
+    pub fn for_field(field: String) -> FastFieldTestCollector {
+        FastFieldTestCollector { field }
     }
 }
 
@@ -212,62 +210,64 @@ impl SegmentCollector for FastFieldSegmentCollector {
     }
 }
 
-// /// Collects in order all of the fast field bytes for all of the
-// /// docs in the `DocSet`
-// ///
-// /// This collector is mainly useful for tests.
-// pub struct BytesFastFieldTestCollector {
-//     field: Field,
-// }
+/// Collects in order all of the fast field bytes for all of the
+/// docs in the `DocSet`
+///
+/// This collector is mainly useful for tests.
+pub struct BytesFastFieldTestCollector {
+    field: Field,
+}
 
-// pub struct BytesFastFieldSegmentCollector {
-//     vals: Vec<u8>,
-//     reader: BytesFastFieldReader,
-// }
+pub struct BytesFastFieldSegmentCollector {
+    vals: Vec<u8>,
+    reader: BytesFastFieldReader,
+}
 
-// impl BytesFastFieldTestCollector {
-//     pub fn for_field(field: Field) -> BytesFastFieldTestCollector {
-//         BytesFastFieldTestCollector { field }
-//     }
-// }
+impl BytesFastFieldTestCollector {
+    pub fn for_field(field: Field) -> BytesFastFieldTestCollector {
+        BytesFastFieldTestCollector { field }
+    }
+}
 
-// impl Collector for BytesFastFieldTestCollector {
-//     type Fruit = Vec<u8>;
-//     type Child = BytesFastFieldSegmentCollector;
+impl Collector for BytesFastFieldTestCollector {
+    type Fruit = Vec<u8>;
+    type Child = BytesFastFieldSegmentCollector;
 
-//     fn for_segment(
-//         &self,
-//         _segment_local_id: u32,
-//         segment_reader: &SegmentReader,
-//     ) -> crate::Result<BytesFastFieldSegmentCollector> {
-//         let reader = segment_reader.fast_fields().bytes(self.field)?;
-//         Ok(BytesFastFieldSegmentCollector {
-//             vals: Vec::new(),
-//             reader,
-//         })
-//     }
+    fn for_segment(
+        &self,
+        _segment_local_id: u32,
+        segment_reader: &SegmentReader,
+    ) -> crate::Result<BytesFastFieldSegmentCollector> {
+        let reader = segment_reader
+            .fast_fields()
+            .bytes(segment_reader.schema().get_field_name(self.field))?;
+        Ok(BytesFastFieldSegmentCollector {
+            vals: Vec::new(),
+            reader,
+        })
+    }
 
-//     fn requires_scoring(&self) -> bool {
-//         false
-//     }
+    fn requires_scoring(&self) -> bool {
+        false
+    }
 
-//     fn merge_fruits(&self, children: Vec<Vec<u8>>) -> crate::Result<Vec<u8>> {
-//         Ok(children.into_iter().flat_map(|c| c.into_iter()).collect())
-//     }
-// }
+    fn merge_fruits(&self, children: Vec<Vec<u8>>) -> crate::Result<Vec<u8>> {
+        Ok(children.into_iter().flat_map(|c| c.into_iter()).collect())
+    }
+}
 
-// impl SegmentCollector for BytesFastFieldSegmentCollector {
-//     type Fruit = Vec<u8>;
+impl SegmentCollector for BytesFastFieldSegmentCollector {
+    type Fruit = Vec<u8>;
 
-//     fn collect(&mut self, doc: u32, _score: Score) {
-//         let data = self.reader.get_bytes(doc);
-//         self.vals.extend(data);
-//     }
+    fn collect(&mut self, doc: u32, _score: Score) {
+        let data = self.reader.get_bytes(doc);
+        self.vals.extend(data);
+    }
 
-//     fn harvest(self) -> <Self as SegmentCollector>::Fruit {
-//         self.vals
-//     }
-// }
+    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
+        self.vals
+    }
+}
 
 fn make_test_searcher() -> crate::Result<Searcher> {
     let schema = Schema::builder().build();

src/collector/top_score_collector.rs

@@ -12,7 +12,7 @@ use crate::collector::tweak_score_top_collector::TweakedScoreTopCollector;
 use crate::collector::{
     CustomScorer, CustomSegmentScorer, ScoreSegmentTweaker, ScoreTweaker, SegmentCollector,
 };
-use crate::fastfield::{FastFieldNotAvailableError, FastValue};
+use crate::fastfield::FastValue;
 use crate::query::Weight;
 use crate::schema::Field;
 use crate::{DocAddress, DocId, Score, SegmentOrdinal, SegmentReader, TantivyError};
@@ -22,7 +22,7 @@ struct FastFieldConvertCollector<
     TFastValue: FastValue,
 > {
     pub collector: TCollector,
-    pub field: String,
+    pub field: Field,
     pub fast_value: std::marker::PhantomData<TFastValue>,
 }
 
@@ -41,8 +41,7 @@ where
         segment: &SegmentReader,
     ) -> crate::Result<Self::Child> {
         let schema = segment.schema();
-        let field = schema.get_field(&self.field)?;
-        let field_entry = schema.get_field_entry(field);
+        let field_entry = schema.get_field_entry(self.field);
         if !field_entry.is_fast() {
             return Err(TantivyError::SchemaError(format!(
                 "Field {:?} is not a fast field.",
@@ -133,17 +132,17 @@ impl fmt::Debug for TopDocs {
 }
 
 struct ScorerByFastFieldReader {
-    sort_column: Arc<dyn Column<u64>>,
+    ff_reader: Arc<dyn Column<u64>>,
 }
 
 impl CustomSegmentScorer<u64> for ScorerByFastFieldReader {
     fn score(&mut self, doc: DocId) -> u64 {
-        self.sort_column.get_val(doc)
+        self.ff_reader.get_val(doc)
     }
 }
 
 struct ScorerByField {
-    field: String,
+    field: Field,
 }
 
 impl CustomScorer<u64> for ScorerByField {
@@ -155,13 +154,10 @@ impl CustomScorer<u64> for ScorerByField {
         // mapping is monotonic, so it is sufficient to compute our top-K docs.
         //
         // The conversion will then happen only on the top-K docs.
-        let sort_column_opt = segment_reader.fast_fields().u64_lenient(&self.field)?;
-        let sort_column = sort_column_opt
-            .ok_or_else(|| FastFieldNotAvailableError {
-                field_name: self.field.clone(),
-            })?
-            .first_or_default_col(0u64);
-        Ok(ScorerByFastFieldReader { sort_column })
+        let ff_reader = segment_reader
+            .fast_fields()
+            .typed_fast_field_reader(segment_reader.schema().get_field_name(self.field))?;
+        Ok(ScorerByFastFieldReader { ff_reader })
     }
 }
 
@@ -294,14 +290,9 @@ impl TopDocs {
     /// the [.order_by_fast_field(...)](TopDocs::order_by_fast_field) method.
     pub fn order_by_u64_field(
         self,
-        field: impl ToString,
+        field: Field,
     ) -> impl Collector<Fruit = Vec<(u64, DocAddress)>> {
-        CustomScoreTopCollector::new(
-            ScorerByField {
-                field: field.to_string(),
-            },
-            self.0.into_tscore(),
-        )
+        CustomScoreTopCollector::new(ScorerByField { field }, self.0.into_tscore())
     }
 
     /// Set top-K to rank documents by a given fast field.
@@ -376,15 +367,15 @@ impl TopDocs {
     /// ```
     pub fn order_by_fast_field<TFastValue>(
         self,
-        fast_field: impl ToString,
+        fast_field: Field,
     ) -> impl Collector<Fruit = Vec<(TFastValue, DocAddress)>>
     where
         TFastValue: FastValue,
     {
-        let u64_collector = self.order_by_u64_field(fast_field.to_string());
+        let u64_collector = self.order_by_u64_field(fast_field);
         FastFieldConvertCollector {
             collector: u64_collector,
-            field: fast_field.to_string(),
+            field: fast_field,
             fast_value: PhantomData,
         }
     }
@@ -886,7 +877,7 @@ mod tests {
         });
         let searcher = index.reader()?.searcher();
 
-        let top_collector = TopDocs::with_limit(4).order_by_u64_field(SIZE);
+        let top_collector = TopDocs::with_limit(4).order_by_u64_field(size);
         let top_docs: Vec<(u64, DocAddress)> = searcher.search(&query, &top_collector)?;
         assert_eq!(
             &top_docs[..],
@@ -925,7 +916,7 @@ mod tests {
         ))?;
         index_writer.commit()?;
         let searcher = index.reader()?.searcher();
-        let top_collector = TopDocs::with_limit(3).order_by_fast_field("birthday");
+        let top_collector = TopDocs::with_limit(3).order_by_fast_field(birthday);
         let top_docs: Vec<(DateTime, DocAddress)> = searcher.search(&AllQuery, &top_collector)?;
         assert_eq!(
             &top_docs[..],
@@ -955,7 +946,7 @@ mod tests {
         ))?;
         index_writer.commit()?;
         let searcher = index.reader()?.searcher();
-        let top_collector = TopDocs::with_limit(3).order_by_fast_field("altitude");
+        let top_collector = TopDocs::with_limit(3).order_by_fast_field(altitude);
         let top_docs: Vec<(i64, DocAddress)> = searcher.search(&AllQuery, &top_collector)?;
         assert_eq!(
             &top_docs[..],
@@ -985,7 +976,7 @@ mod tests {
         ))?;
         index_writer.commit()?;
         let searcher = index.reader()?.searcher();
-        let top_collector = TopDocs::with_limit(3).order_by_fast_field("altitude");
+        let top_collector = TopDocs::with_limit(3).order_by_fast_field(altitude);
         let top_docs: Vec<(f64, DocAddress)> = searcher.search(&AllQuery, &top_collector)?;
         assert_eq!(
             &top_docs[..],
@@ -1013,7 +1004,7 @@ mod tests {
                 .unwrap();
         });
         let searcher = index.reader().unwrap().searcher();
-        let top_collector = TopDocs::with_limit(4).order_by_u64_field("missing_field");
+        let top_collector = TopDocs::with_limit(4).order_by_u64_field(Field::from_field_id(2));
         let segment_reader = searcher.segment_reader(0u32);
         top_collector
             .for_segment(0, segment_reader)
@@ -1031,7 +1022,7 @@ mod tests {
         index_writer.commit()?;
         let searcher = index.reader()?.searcher();
         let segment = searcher.segment_reader(0);
-        let top_collector = TopDocs::with_limit(4).order_by_u64_field(SIZE);
+        let top_collector = TopDocs::with_limit(4).order_by_u64_field(size);
         let err = top_collector.for_segment(0, segment).err().unwrap();
         assert!(matches!(err, crate::TantivyError::SchemaError(_)));
         Ok(())
@@ -1048,7 +1039,7 @@ mod tests {
         index_writer.commit()?;
         let searcher = index.reader()?.searcher();
         let segment = searcher.segment_reader(0);
-        let top_collector = TopDocs::with_limit(4).order_by_fast_field::<i64>(SIZE);
+        let top_collector = TopDocs::with_limit(4).order_by_fast_field::<i64>(size);
         let err = top_collector.for_segment(0, segment).err().unwrap();
         assert!(
             matches!(err, crate::TantivyError::SchemaError(msg) if msg == "Field \"size\" is not a fast field.")

src/core/index.rs

@@ -19,7 +19,7 @@ use crate::error::{DataCorruption, TantivyError};
 use crate::indexer::index_writer::{MAX_NUM_THREAD, MEMORY_ARENA_NUM_BYTES_MIN};
 use crate::indexer::segment_updater::save_metas;
 use crate::reader::{IndexReader, IndexReaderBuilder};
-use crate::schema::{Field, FieldType, Schema};
+use crate::schema::{Cardinality, Field, FieldType, Schema};
 use crate::tokenizer::{TextAnalyzer, TokenizerManager};
 use crate::IndexWriter;
 
@@ -93,7 +93,7 @@ fn save_new_metas(
 /// let body_field = schema_builder.add_text_field("body", TEXT);
 /// let number_field = schema_builder.add_u64_field(
 ///     "number",
-///     NumericOptions::default().set_fast(),
+///     NumericOptions::default().set_fast(Cardinality::SingleValue),
 /// );
 ///
 /// let schema = schema_builder.build();
@@ -245,6 +245,12 @@ impl IndexBuilder {
                         sort_by_field.field
                     )));
                 }
+                if entry.field_type().fastfield_cardinality() != Some(Cardinality::SingleValue) {
+                    return Err(TantivyError::InvalidArgument(format!(
+                        "Only single value fast field Cardinality supported for sorting index {}",
+                        sort_by_field.field
+                    )));
+                }
             }
             Ok(())
         } else {

src/core/segment_reader.rs

@@ -7,7 +7,7 @@ use fail::fail_point;
 use crate::core::{InvertedIndexReader, Segment, SegmentComponent, SegmentId};
 use crate::directory::{CompositeFile, FileSlice};
 use crate::error::DataCorruption;
-use crate::fastfield::{intersect_alive_bitsets, AliveBitSet, FastFieldReaders};
+use crate::fastfield::{intersect_alive_bitsets, AliveBitSet, FacetReader, FastFieldReaders};
 use crate::fieldnorm::{FieldNormReader, FieldNormReaders};
 use crate::schema::{Field, FieldType, IndexRecordOption, Schema};
 use crate::space_usage::SegmentSpaceUsage;
@@ -90,8 +90,25 @@ impl SegmentReader {
     }
 
     /// Accessor to the `FacetReader` associated with a given `Field`.
-    pub fn facet_reader(&self, field: Field) -> crate::Result<()> {
-        todo!();
+    pub fn facet_reader(&self, field: Field) -> crate::Result<FacetReader> {
+        let field_entry = self.schema.get_field_entry(field);
+
+        match field_entry.field_type() {
+            FieldType::Facet(_) => {
+                let term_ords_reader =
+                    self.fast_fields().u64s(self.schema.get_field_name(field))?;
+                let termdict = self
+                    .termdict_composite
+                    .open_read(field)
+                    .map(TermDictionary::open)
+                    .unwrap_or_else(|| Ok(TermDictionary::empty()))?;
+                Ok(FacetReader::new(term_ords_reader, termdict))
+            }
+            _ => Err(crate::TantivyError::InvalidArgument(format!(
+                "Field {:?} is not a facet field.",
+                field_entry.name()
+            ))),
+        }
     }
 
     /// Accessor to the segment's `Field norms`'s reader.
@@ -153,7 +170,9 @@ impl SegmentReader {
         let schema = segment.schema();
 
         let fast_fields_data = segment.open_read(SegmentComponent::FastFields)?;
-        let fast_fields_readers = Arc::new(FastFieldReaders::open(fast_fields_data)?);
+        let fast_fields_composite = CompositeFile::open(&fast_fields_data)?;
+        let fast_fields_readers =
+            Arc::new(FastFieldReaders::new(schema.clone(), fast_fields_composite));
         let fieldnorm_data = segment.open_read(SegmentComponent::FieldNorms)?;
         let fieldnorm_readers = FieldNormReaders::open(fieldnorm_data)?;
 

src/fastfield/error.rs

@@ -8,7 +8,7 @@ use crate::schema::FieldEntry;
 #[derive(Debug, Error)]
 #[error("Fast field not available: '{field_name:?}'")]
 pub struct FastFieldNotAvailableError {
-    pub(crate) field_name: String,
+    field_name: String,
 }
 
 impl FastFieldNotAvailableError {

src/fastfield/multivalued/mod.rs

@@ -38,7 +38,7 @@ mod tests {
         let mut schema_builder = Schema::builder();
         let field = schema_builder.add_u64_field(
             "multifield",
-            NumericOptions::default().set_fast(),
+            NumericOptions::default().set_fast(Cardinality::MultiValues),
         );
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -74,7 +74,7 @@ mod tests {
         let date_field = schema_builder.add_date_field(
             "multi_date_field",
             DateOptions::default()
-                .set_fast()
+                .set_fast(Cardinality::MultiValues)
                 .set_indexed()
                 .set_fieldnorm()
                 .set_stored(),
@@ -215,7 +215,7 @@ mod tests {
         let mut schema_builder = Schema::builder();
         let field = schema_builder.add_i64_field(
             "multifield",
-            NumericOptions::default().set_fast(),
+            NumericOptions::default().set_fast(Cardinality::MultiValues),
         );
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -246,7 +246,7 @@ mod tests {
         let mut schema_builder = Schema::builder();
         let bool_field = schema_builder.add_bool_field(
             "multifield",
-            NumericOptions::default().set_fast(),
+            NumericOptions::default().set_fast(Cardinality::MultiValues),
         );
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -278,7 +278,7 @@ mod tests {
         let field = schema_builder.add_u64_field(
             "multifield",
             NumericOptions::default()
-                .set_fast()
+                .set_fast(Cardinality::MultiValues)
                 .set_indexed(),
         );
         let schema = schema_builder.build();
@@ -424,7 +424,7 @@ mod bench {
         let mut builder = crate::schema::SchemaBuilder::new();
 
         let fast_multi =
-            crate::schema::NumericOptions::default().set_fast();
+            crate::schema::NumericOptions::default().set_fast(Cardinality::MultiValues);
         let multi_field = builder.add_f64_field("f64s", fast_multi);
 
         let index = crate::Index::create_in_ram(builder.build());
@@ -504,7 +504,7 @@ mod bench {
         let path = Path::new("test");
         let directory: RamDirectory = RamDirectory::create();
         let field = {
-            let options = NumericOptions::default().set_fast();
+            let options = NumericOptions::default().set_fast(Cardinality::MultiValues);
             let mut schema_builder = Schema::builder();
             let field = schema_builder.add_u64_field("field", options);
             let schema = schema_builder.build();
@@ -562,7 +562,7 @@ mod bench {
 
         b.iter(|| {
             let directory: RamDirectory = RamDirectory::create();
-            let options = NumericOptions::default().set_fast();
+            let options = NumericOptions::default().set_fast(Cardinality::MultiValues);
             let mut schema_builder = Schema::builder();
             let field = schema_builder.add_u64_field("field", options);
             let schema = schema_builder.build();
@@ -595,7 +595,7 @@ mod bench {
 
         b.iter(|| {
             let directory: RamDirectory = RamDirectory::create();
-            let options = NumericOptions::default().set_fast();
+            let options = NumericOptions::default().set_fast(Cardinality::MultiValues);
             let mut schema_builder = Schema::builder();
             let field = schema_builder.add_u64_field("field", options);
             let schema = schema_builder.build();

src/fastfield/multivalued/reader.rs

@@ -137,7 +137,7 @@ mod tests {
         let date_field = schema_builder.add_date_field(
             "multi_date_field",
             DateOptions::default()
-                .set_fast()
+                .set_fast(Cardinality::MultiValues)
                 .set_indexed()
                 .set_fieldnorm()
                 .set_precision(DatePrecision::Microseconds)
@@ -188,7 +188,7 @@ mod tests {
         let date_field = schema_builder.add_date_field(
             "multi_date_field",
             DateOptions::default()
-                .set_fast()
+                .set_fast(Cardinality::MultiValues)
                 // TODO: Test different precision after fixing https://github.com/quickwit-oss/tantivy/issues/1783
                 .set_precision(DatePrecision::Microseconds)
                 .set_indexed()
@@ -307,7 +307,7 @@ mod tests {
         let mut schema_builder = Schema::builder();
         let field_options = NumericOptions::default()
             .set_indexed()
-            .set_fast();
+            .set_fast(Cardinality::MultiValues);
         let item_field = schema_builder.add_i64_field("items", field_options);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);

src/fastfield/readers.rs

@@ -1,16 +1,12 @@
-use std::io;
 use std::net::Ipv6Addr;
 use std::sync::Arc;
 
-use columnar::{
-    ColumnType, ColumnValues, ColumnarReader, DynamicColumn, DynamicColumnHandle,
-    HasAssociatedColumnType, NumericalType,
-};
 use fastfield_codecs::{open, open_u128, Column};
 
+use super::multivalued::MultiValuedFastFieldReader;
 use crate::directory::{CompositeFile, FileSlice};
-use crate::fastfield::{FastFieldNotAvailableError, FastValue};
-use crate::schema::{Field, FieldType, Schema};
+use crate::fastfield::{BytesFastFieldReader, FastFieldNotAvailableError, FastValue};
+use crate::schema::{Cardinality, Field, FieldType, Schema};
 use crate::space_usage::PerFieldSpaceUsage;
 use crate::{DateTime, TantivyError};
 
@@ -20,152 +16,315 @@ use crate::{DateTime, TantivyError};
 /// and just wraps several `HashMap`.
 #[derive(Clone)]
 pub struct FastFieldReaders {
-    columnar: Arc<ColumnarReader>,
+    schema: Schema,
+    fast_fields_composite: CompositeFile,
+}
+#[derive(Eq, PartialEq, Debug)]
+pub(crate) enum FastType {
+    I64,
+    U64,
+    U128,
+    F64,
+    Bool,
+    Date,
+}
+
+pub(crate) fn type_and_cardinality(field_type: &FieldType) -> Option<(FastType, Cardinality)> {
+    match field_type {
+        FieldType::U64(options) => options
+            .get_fastfield_cardinality()
+            .map(|cardinality| (FastType::U64, cardinality)),
+        FieldType::I64(options) => options
+            .get_fastfield_cardinality()
+            .map(|cardinality| (FastType::I64, cardinality)),
+        FieldType::F64(options) => options
+            .get_fastfield_cardinality()
+            .map(|cardinality| (FastType::F64, cardinality)),
+        FieldType::Bool(options) => options
+            .get_fastfield_cardinality()
+            .map(|cardinality| (FastType::Bool, cardinality)),
+        FieldType::Date(options) => options
+            .get_fastfield_cardinality()
+            .map(|cardinality| (FastType::Date, cardinality)),
+        FieldType::Facet(_) => Some((FastType::U64, Cardinality::MultiValues)),
+        FieldType::Str(options) if options.is_fast() => {
+            Some((FastType::U64, Cardinality::MultiValues))
+        }
+        FieldType::IpAddr(options) => options
+            .get_fastfield_cardinality()
+            .map(|cardinality| (FastType::U128, cardinality)),
+        _ => None,
+    }
 }
 
 impl FastFieldReaders {
-    pub(crate) fn open(fast_field_file: FileSlice) -> io::Result<FastFieldReaders> {
-        let columnar = Arc::new(ColumnarReader::open(fast_field_file)?);
-        Ok(FastFieldReaders { columnar })
+    pub(crate) fn new(schema: Schema, fast_fields_composite: CompositeFile) -> FastFieldReaders {
+        FastFieldReaders {
+            schema,
+            fast_fields_composite,
+        }
     }
 
     pub(crate) fn space_usage(&self) -> PerFieldSpaceUsage {
-        todo!()
+        self.fast_fields_composite.space_usage()
     }
 
-    pub fn typed_column_opt<T>(
+    #[doc(hidden)]
+    pub fn fast_field_data(&self, field: Field, idx: usize) -> crate::Result<FileSlice> {
+        self.fast_fields_composite
+            .open_read_with_idx(field, idx)
+            .ok_or_else(|| {
+                let field_name = self.schema.get_field_entry(field).name();
+                TantivyError::SchemaError(format!("Field({}) data was not found", field_name))
+            })
+    }
+
+    fn check_type(
+        &self,
+        field: Field,
+        expected_fast_type: FastType,
+        expected_cardinality: Cardinality,
+    ) -> crate::Result<()> {
+        let field_entry = self.schema.get_field_entry(field);
+        let (fast_type, cardinality) =
+            type_and_cardinality(field_entry.field_type()).ok_or_else(|| {
+                crate::TantivyError::SchemaError(format!(
+                    "Field {:?} is not a fast field.",
+                    field_entry.name()
+                ))
+            })?;
+        if fast_type != expected_fast_type {
+            return Err(crate::TantivyError::SchemaError(format!(
+                "Field {:?} is of type {:?}, expected {:?}.",
+                field_entry.name(),
+                fast_type,
+                expected_fast_type
+            )));
+        }
+        if cardinality != expected_cardinality {
+            return Err(crate::TantivyError::SchemaError(format!(
+                "Field {:?} is of cardinality {:?}, expected {:?}.",
+                field_entry.name(),
+                cardinality,
+                expected_cardinality
+            )));
+        }
+        Ok(())
+    }
+
+    pub(crate) fn typed_fast_field_reader_with_idx<TFastValue: FastValue>(
         &self,
         field_name: &str,
-    ) -> crate::Result<Option<columnar::Column<T>>>
-    where
-        T: PartialOrd + Copy + HasAssociatedColumnType + Send + Sync + Default + 'static,
-        DynamicColumn: Into<Option<columnar::Column<T>>>,
-    {
-        let column_type = T::column_type();
-        let Some(dynamic_column_handle) = self.column_handle(field_name, column_type)?
-        else {
-            return Ok(None);
-        };
-        let dynamic_column = dynamic_column_handle.open()?;
-        Ok(dynamic_column.into())
+        index: usize,
+    ) -> crate::Result<Arc<dyn Column<TFastValue>>> {
+        let field = self.schema.get_field(field_name)?;
+
+        let fast_field_slice = self.fast_field_data(field, index)?;
+        let bytes = fast_field_slice.read_bytes()?;
+        let column = fastfield_codecs::open(bytes)?;
+        Ok(column)
     }
 
-    pub fn column_num_bytes(&self, field: &str) -> crate::Result<usize> {
-        Ok(self
-            .columnar
-            .read_columns(field)?
-            .into_iter()
-            .map(|column_handle| column_handle.num_bytes())
-            .sum())
+    pub(crate) fn typed_fast_field_reader<TFastValue: FastValue>(
+        &self,
+        field_name: &str,
+    ) -> crate::Result<Arc<dyn Column<TFastValue>>> {
+        self.typed_fast_field_reader_with_idx(field_name, 0)
     }
 
-    pub fn typed_column_first_or_default<T>(&self, field: &str) -> crate::Result<Arc<dyn Column<T>>>
-    where
-        T: PartialOrd + Copy + HasAssociatedColumnType + Send + Sync + Default + 'static,
-        DynamicColumn: Into<Option<columnar::Column<T>>>,
-    {
-        let col_opt = self.typed_column_opt(field)?;
-        if let Some(col) = col_opt {
-            Ok(col.first_or_default_col(T::default()))
-        } else {
-            todo!();
-        }
+    pub(crate) fn typed_fast_field_multi_reader<TFastValue: FastValue>(
+        &self,
+        field_name: &str,
+    ) -> crate::Result<MultiValuedFastFieldReader<TFastValue>> {
+        let idx_reader = self.typed_fast_field_reader(field_name)?;
+        let vals_reader = self.typed_fast_field_reader_with_idx(field_name, 1)?;
+        Ok(MultiValuedFastFieldReader::open(idx_reader, vals_reader))
     }
 
     /// Returns the `u64` fast field reader reader associated with `field`.
     ///
     /// If `field` is not a u64 fast field, this method returns an Error.
-    pub fn u64(&self, field: &str) -> crate::Result<Arc<dyn ColumnValues<u64>>> {
-        self.typed_column_first_or_default(field)
-    }
-
-    /// Returns the `date` fast field reader reader associated with `field`.
-    ///
-    /// If `field` is not a date fast field, this method returns an Error.
-    pub fn date(&self, field: &str) -> crate::Result<Arc<dyn ColumnValues<columnar::DateTime>>> {
-        self.typed_column_first_or_default(field)
+    pub fn u64(&self, field_name: &str) -> crate::Result<Arc<dyn Column<u64>>> {
+        self.check_type(
+            self.schema.get_field(field_name)?,
+            FastType::U64,
+            Cardinality::SingleValue,
+        )?;
+        self.typed_fast_field_reader(field_name)
     }
 
     /// Returns the `ip` fast field reader reader associated to `field`.
     ///
     /// If `field` is not a u128 fast field, this method returns an Error.
-    pub fn ip_addr(&self, field: &str) -> crate::Result<Arc<dyn Column<Ipv6Addr>>> {
-        todo!();
-        // self.check_type(field, FastType::U128)?;
-        // let bytes = self.fast_field_data(field, 0)?.read_bytes()?;
-        // Ok(open_u128::<Ipv6Addr>(bytes)?)
+    pub fn ip_addr(&self, field_name: &str) -> crate::Result<Arc<dyn Column<Ipv6Addr>>> {
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::U128, Cardinality::SingleValue)?;
+        let bytes = self.fast_field_data(field, 0)?.read_bytes()?;
+        Ok(open_u128::<Ipv6Addr>(bytes)?)
+    }
+
+    /// Returns the `ip` fast field reader reader associated to `field`.
+    ///
+    /// If `field` is not a u128 fast field, this method returns an Error.
+    pub fn ip_addrs(
+        &self,
+        field_name: &str,
+    ) -> crate::Result<MultiValuedFastFieldReader<Ipv6Addr>> {
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::U128, Cardinality::MultiValues)?;
+        let idx_reader: Arc<dyn Column<u64>> = self.typed_fast_field_reader(field_name)?;
+
+        let bytes = self.fast_field_data(field, 1)?.read_bytes()?;
+        let vals_reader = open_u128::<Ipv6Addr>(bytes)?;
+
+        Ok(MultiValuedFastFieldReader::open(idx_reader, vals_reader))
     }
 
     /// Returns the `u128` fast field reader reader associated to `field`.
     ///
     /// If `field` is not a u128 fast field, this method returns an Error.
-    pub(crate) fn u128(&self, field: &str) -> crate::Result<Arc<dyn Column<u128>>> {
-        todo!();
+    pub(crate) fn u128(&self, field_name: &str) -> crate::Result<Arc<dyn Column<u128>>> {
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::U128, Cardinality::SingleValue)?;
+        let bytes = self.fast_field_data(field, 0)?.read_bytes()?;
+        Ok(open_u128::<u128>(bytes)?)
     }
 
-    pub fn column_handle(
-        &self,
-        field_name: &str,
-        column_type: ColumnType,
-    ) -> crate::Result<Option<DynamicColumnHandle>> {
-        let dynamic_column_handle_opt = self
-            .columnar
-            .read_columns(field_name)?
-            .into_iter()
-            .filter(|column| column.column_type() == column_type)
-            .next();
-        Ok(dynamic_column_handle_opt)
+    /// Returns the `u128` multi-valued fast field reader reader associated to `field`.
+    ///
+    /// If `field` is not a u128 multi-valued fast field, this method returns an Error.
+    pub fn u128s(&self, field_name: &str) -> crate::Result<MultiValuedFastFieldReader<u128>> {
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::U128, Cardinality::MultiValues)?;
+        let idx_reader: Arc<dyn Column<u64>> =
+            self.typed_fast_field_reader(self.schema.get_field_name(field))?;
+
+        let bytes = self.fast_field_data(field, 1)?.read_bytes()?;
+        let vals_reader = open_u128::<u128>(bytes)?;
+
+        Ok(MultiValuedFastFieldReader::open(idx_reader, vals_reader))
     }
 
-    pub fn u64_lenient(&self, field_name: &str) -> crate::Result<Option<columnar::Column<u64>>> {
-        for col in self.columnar.read_columns(field_name)? {
-            if let Some(col_u64) = col.open_u64_lenient()? {
-                return Ok(Some(col_u64));
-            }
-        }
-        Ok(None)
+    /// Returns the `u64` fast field reader reader associated with `field`, regardless of whether
+    /// the given field is effectively of type `u64` or not.
+    ///
+    /// If not, the fastfield reader will returns the u64-value associated with the original
+    /// FastValue.
+    pub fn u64_lenient(&self, field_name: &str) -> crate::Result<Arc<dyn Column<u64>>> {
+        self.typed_fast_field_reader(field_name)
     }
 
     /// Returns the `i64` fast field reader reader associated with `field`.
     ///
     /// If `field` is not a i64 fast field, this method returns an Error.
     pub fn i64(&self, field_name: &str) -> crate::Result<Arc<dyn Column<i64>>> {
-        self.typed_column_first_or_default(field_name)
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::I64, Cardinality::SingleValue)?;
+        self.typed_fast_field_reader(self.schema.get_field_name(field))
+    }
+
+    /// Returns the `date` fast field reader reader associated with `field`.
+    ///
+    /// If `field` is not a date fast field, this method returns an Error.
+    pub fn date(&self, field_name: &str) -> crate::Result<Arc<dyn Column<DateTime>>> {
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::Date, Cardinality::SingleValue)?;
+        self.typed_fast_field_reader(field_name)
     }
 
     /// Returns the `f64` fast field reader reader associated with `field`.
     ///
     /// If `field` is not a f64 fast field, this method returns an Error.
     pub fn f64(&self, field_name: &str) -> crate::Result<Arc<dyn Column<f64>>> {
-        self.typed_column_first_or_default(field_name)
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::F64, Cardinality::SingleValue)?;
+        self.typed_fast_field_reader(field_name)
     }
 
     /// Returns the `bool` fast field reader reader associated with `field`.
     ///
     /// If `field` is not a bool fast field, this method returns an Error.
     pub fn bool(&self, field_name: &str) -> crate::Result<Arc<dyn Column<bool>>> {
-        self.typed_column_first_or_default(field_name)
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::Bool, Cardinality::SingleValue)?;
+        self.typed_fast_field_reader(field_name)
     }
 
-    // Returns the `bytes` fast field reader associated with `field`.
-    //
-    // If `field` is not a bytes fast field, returns an Error.
-    // pub fn bytes(&self, field: Field) -> crate::Result<BytesFastFieldReader> {
-    // let field_entry = self.schema.get_field_entry(field);
-    // if let FieldType::Bytes(bytes_option) = field_entry.field_type() {
-    //     if !bytes_option.is_fast() {
-    //         return Err(crate::TantivyError::SchemaError(format!(
-    //             "Field {:?} is not a fast field.",
-    //             field_entry.name()
-    //         )));
-    //     }
-    //     let fast_field_idx_file = self.fast_field_data(field, 0)?;
-    //     let fast_field_idx_bytes = fast_field_idx_file.read_bytes()?;
-    //     let idx_reader = open(fast_field_idx_bytes)?;
-    //     let data = self.fast_field_data(field, 1)?;
-    //     BytesFastFieldReader::open(idx_reader, data)
-    // } else {
-    //     Err(FastFieldNotAvailableError::new(field_entry).into())
-    // }
-    // }
+    /// Returns a `u64s` multi-valued fast field reader reader associated with `field`.
+    ///
+    /// If `field` is not a u64 multi-valued fast field, this method returns an Error.
+    pub fn u64s(&self, field_name: &str) -> crate::Result<MultiValuedFastFieldReader<u64>> {
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::U64, Cardinality::MultiValues)?;
+        self.typed_fast_field_multi_reader(field_name)
+    }
+
+    /// Returns a `u64s` multi-valued fast field reader reader associated with `field`, regardless
+    /// of whether the given field is effectively of type `u64` or not.
+    ///
+    /// If `field` is not a u64 multi-valued fast field, this method returns an Error.
+    pub fn u64s_lenient(&self, field_name: &str) -> crate::Result<MultiValuedFastFieldReader<u64>> {
+        self.typed_fast_field_multi_reader(field_name)
+    }
+
+    /// Returns a `i64s` multi-valued fast field reader reader associated with `field`.
+    ///
+    /// If `field` is not a i64 multi-valued fast field, this method returns an Error.
+    pub fn i64s(&self, field_name: &str) -> crate::Result<MultiValuedFastFieldReader<i64>> {
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::I64, Cardinality::MultiValues)?;
+        self.typed_fast_field_multi_reader(self.schema.get_field_name(field))
+    }
+
+    /// Returns a `f64s` multi-valued fast field reader reader associated with `field`.
+    ///
+    /// If `field` is not a f64 multi-valued fast field, this method returns an Error.
+    pub fn f64s(&self, field_name: &str) -> crate::Result<MultiValuedFastFieldReader<f64>> {
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::F64, Cardinality::MultiValues)?;
+        self.typed_fast_field_multi_reader(self.schema.get_field_name(field))
+    }
+
+    /// Returns a `bools` multi-valued fast field reader reader associated with `field`.
+    ///
+    /// If `field` is not a bool multi-valued fast field, this method returns an Error.
+    pub fn bools(&self, field_name: &str) -> crate::Result<MultiValuedFastFieldReader<bool>> {
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::Bool, Cardinality::MultiValues)?;
+        self.typed_fast_field_multi_reader(self.schema.get_field_name(field))
+    }
+
+    /// Returns a `time::OffsetDateTime` multi-valued fast field reader reader associated with
+    /// `field`.
+    ///
+    /// If `field` is not a `time::OffsetDateTime` multi-valued fast field, this method returns an
+    /// Error.
+    pub fn dates(&self, field_name: &str) -> crate::Result<MultiValuedFastFieldReader<DateTime>> {
+        let field = self.schema.get_field(field_name)?;
+        self.check_type(field, FastType::Date, Cardinality::MultiValues)?;
+        self.typed_fast_field_multi_reader(self.schema.get_field_name(field))
+    }
+
+    /// Returns the `bytes` fast field reader associated with `field`.
+    ///
+    /// If `field` is not a bytes fast field, returns an Error.
+    pub fn bytes(&self, field_name: &str) -> crate::Result<BytesFastFieldReader> {
+        let field = self.schema.get_field(field_name)?;
+        let field_entry = self.schema.get_field_entry(field);
+        if let FieldType::Bytes(bytes_option) = field_entry.field_type() {
+            if !bytes_option.is_fast() {
+                return Err(crate::TantivyError::SchemaError(format!(
+                    "Field {:?} is not a fast field.",
+                    field_entry.name()
+                )));
+            }
+            let fast_field_idx_file = self.fast_field_data(field, 0)?;
+            let fast_field_idx_bytes = fast_field_idx_file.read_bytes()?;
+            let idx_reader = open(fast_field_idx_bytes)?;
+            let data = self.fast_field_data(field, 1)?;
+            BytesFastFieldReader::open(idx_reader, data)
+        } else {
+            Err(FastFieldNotAvailableError::new(field_entry).into())
+        }
+    }
 }

src/fastfield/writer.rs

@@ -1,141 +1,558 @@
 use std::collections::HashMap;
 use std::io;
 
-use columnar::{ColumnType, ColumnarWriter, NumericalType, NumericalValue};
 use common;
 use fastfield_codecs::{Column, MonotonicallyMappableToU128, MonotonicallyMappableToU64};
 use rustc_hash::FxHashMap;
 use tantivy_bitpacker::BlockedBitpacker;
 
+use super::multivalued::{MultiValueU128FastFieldWriter, MultiValuedFastFieldWriter};
 use super::FastFieldType;
-use crate::fastfield::CompositeFastFieldSerializer;
+use crate::fastfield::{BytesFastFieldWriter, CompositeFastFieldSerializer};
 use crate::indexer::doc_id_mapping::DocIdMapping;
 use crate::postings::UnorderedTermId;
-use crate::schema::{Document, Field, FieldEntry, FieldType, Schema, Type, Value};
+use crate::schema::{Cardinality, Document, Field, FieldEntry, FieldType, Schema, Value};
 use crate::termdict::TermOrdinal;
-use crate::{DatePrecision, DocId};
+use crate::DatePrecision;
 
 /// The `FastFieldsWriter` groups all of the fast field writers.
 pub struct FastFieldsWriter {
-    columnar_writer: ColumnarWriter,
-    fast_field_names: Vec<Option<String>>, //< TODO see if we can cash the field name hash too.
-    date_precisions: Vec<DatePrecision>,
-    num_docs: DocId,
+    term_id_writers: Vec<MultiValuedFastFieldWriter>,
+    single_value_writers: Vec<IntFastFieldWriter>,
+    u128_value_writers: Vec<U128FastFieldWriter>,
+    u128_multi_value_writers: Vec<MultiValueU128FastFieldWriter>,
+    multi_values_writers: Vec<MultiValuedFastFieldWriter>,
+    bytes_value_writers: Vec<BytesFastFieldWriter>,
+}
+
+pub(crate) fn unexpected_value(expected: &str, actual: &Value) -> crate::TantivyError {
+    crate::TantivyError::SchemaError(format!(
+        "Expected a {:?} in fast field, but got {:?}",
+        expected, actual
+    ))
+}
+
+fn fast_field_default_value(field_entry: &FieldEntry) -> u64 {
+    match *field_entry.field_type() {
+        FieldType::I64(_) | FieldType::Date(_) => common::i64_to_u64(0i64),
+        FieldType::F64(_) => common::f64_to_u64(0.0f64),
+        _ => 0u64,
+    }
 }
 
 impl FastFieldsWriter {
     /// Create all `FastFieldWriter` required by the schema.
     pub fn from_schema(schema: &Schema) -> FastFieldsWriter {
-        let mut columnar_writer = ColumnarWriter::default();
-        let mut fast_fields: Vec<Option<String>> = vec![None; schema.num_fields()];
-        let mut date_precisions: Vec<DatePrecision> =
-            std::iter::repeat_with(DatePrecision::default)
-                .take(schema.num_fields())
-                .collect();
-        // TODO see other types
-        for (field_id, field_entry) in schema.fields() {
-            if !field_entry.field_type().is_fast() {
-                continue;
-            }
-            fast_fields[field_id.field_id() as usize] = Some(field_entry.name().to_string());
-            let column_type = match field_entry.field_type().value_type() {
-                Type::Str => ColumnType::Str,
-                Type::U64 => ColumnType::Numerical(NumericalType::U64),
-                Type::I64 => ColumnType::Numerical(NumericalType::I64),
-                Type::F64 => ColumnType::Numerical(NumericalType::F64),
-                Type::Bool => ColumnType::Bool,
-                Type::Date => ColumnType::DateTime,
-                Type::Facet => ColumnType::Str,
-                Type::Bytes => todo!(),
-                Type::Json => {
-                    continue;
+        let mut u128_value_writers = Vec::new();
+        let mut u128_multi_value_writers = Vec::new();
+        let mut single_value_writers = Vec::new();
+        let mut term_id_writers = Vec::new();
+        let mut multi_values_writers = Vec::new();
+        let mut bytes_value_writers = Vec::new();
+
+        for (field, field_entry) in schema.fields() {
+            match field_entry.field_type() {
+                FieldType::I64(ref int_options)
+                | FieldType::U64(ref int_options)
+                | FieldType::F64(ref int_options)
+                | FieldType::Bool(ref int_options) => {
+                    match int_options.get_fastfield_cardinality() {
+                        Some(Cardinality::SingleValue) => {
+                            let mut fast_field_writer = IntFastFieldWriter::new(field, None);
+                            let default_value = fast_field_default_value(field_entry);
+                            fast_field_writer.set_val_if_missing(default_value);
+                            single_value_writers.push(fast_field_writer);
+                        }
+                        Some(Cardinality::MultiValues) => {
+                            let fast_field_writer = MultiValuedFastFieldWriter::new(
+                                field,
+                                FastFieldType::Numeric,
+                                None,
+                            );
+                            multi_values_writers.push(fast_field_writer);
+                        }
+                        None => {}
+                    }
                 }
-                Type::IpAddr => todo!(),
-            };
-            if let FieldType::Date(date_options) = field_entry.field_type() {
-                date_precisions[field_id.field_id() as usize] = date_options.get_precision();
+                FieldType::Date(ref options) => match options.get_fastfield_cardinality() {
+                    Some(Cardinality::SingleValue) => {
+                        let mut fast_field_writer =
+                            IntFastFieldWriter::new(field, Some(options.get_precision()));
+                        let default_value = fast_field_default_value(field_entry);
+                        fast_field_writer.set_val_if_missing(default_value);
+                        single_value_writers.push(fast_field_writer);
+                    }
+                    Some(Cardinality::MultiValues) => {
+                        let fast_field_writer = MultiValuedFastFieldWriter::new(
+                            field,
+                            FastFieldType::Numeric,
+                            Some(options.get_precision()),
+                        );
+                        multi_values_writers.push(fast_field_writer);
+                    }
+                    None => {}
+                },
+                FieldType::Facet(_) => {
+                    let fast_field_writer =
+                        MultiValuedFastFieldWriter::new(field, FastFieldType::Facet, None);
+                    term_id_writers.push(fast_field_writer);
+                }
+                FieldType::Str(_) if field_entry.is_fast() => {
+                    let fast_field_writer =
+                        MultiValuedFastFieldWriter::new(field, FastFieldType::String, None);
+                    term_id_writers.push(fast_field_writer);
+                }
+                FieldType::Bytes(bytes_option) => {
+                    if bytes_option.is_fast() {
+                        let fast_field_writer = BytesFastFieldWriter::new(field);
+                        bytes_value_writers.push(fast_field_writer);
+                    }
+                }
+                FieldType::IpAddr(opt) => {
+                    if opt.is_fast() {
+                        match opt.get_fastfield_cardinality() {
+                            Some(Cardinality::SingleValue) => {
+                                let fast_field_writer = U128FastFieldWriter::new(field);
+                                u128_value_writers.push(fast_field_writer);
+                            }
+                            Some(Cardinality::MultiValues) => {
+                                let fast_field_writer = MultiValueU128FastFieldWriter::new(field);
+                                u128_multi_value_writers.push(fast_field_writer);
+                            }
+                            None => {}
+                        }
+                    }
+                }
+                FieldType::Str(_) | FieldType::JsonObject(_) => {}
             }
-            columnar_writer.record_column_type(field_entry.name(), column_type);
         }
         FastFieldsWriter {
-            columnar_writer,
-            fast_field_names: fast_fields,
-            num_docs: 0u32,
-            date_precisions,
+            u128_value_writers,
+            u128_multi_value_writers,
+            term_id_writers,
+            single_value_writers,
+            multi_values_writers,
+            bytes_value_writers,
         }
     }
 
     /// The memory used (inclusive childs)
     pub fn mem_usage(&self) -> usize {
-        self.columnar_writer.mem_usage()
+        self.term_id_writers
+            .iter()
+            .map(|w| w.mem_usage())
+            .sum::<usize>()
+            + self
+                .single_value_writers
+                .iter()
+                .map(|w| w.mem_usage())
+                .sum::<usize>()
+            + self
+                .multi_values_writers
+                .iter()
+                .map(|w| w.mem_usage())
+                .sum::<usize>()
+            + self
+                .bytes_value_writers
+                .iter()
+                .map(|w| w.mem_usage())
+                .sum::<usize>()
+            + self
+                .u128_value_writers
+                .iter()
+                .map(|w| w.mem_usage())
+                .sum::<usize>()
+            + self
+                .u128_multi_value_writers
+                .iter()
+                .map(|w| w.mem_usage())
+                .sum::<usize>()
     }
 
+    /// Get the `FastFieldWriter` associated with a field.
+    pub fn get_term_id_writer(&self, field: Field) -> Option<&MultiValuedFastFieldWriter> {
+        // TODO optimize
+        self.term_id_writers
+            .iter()
+            .find(|field_writer| field_writer.field() == field)
+    }
+
+    /// Get the `FastFieldWriter` associated with a field.
+    pub fn get_field_writer(&self, field: Field) -> Option<&IntFastFieldWriter> {
+        // TODO optimize
+        self.single_value_writers
+            .iter()
+            .find(|field_writer| field_writer.field() == field)
+    }
+
+    /// Get the `FastFieldWriter` associated with a field.
+    pub fn get_field_writer_mut(&mut self, field: Field) -> Option<&mut IntFastFieldWriter> {
+        // TODO optimize
+        self.single_value_writers
+            .iter_mut()
+            .find(|field_writer| field_writer.field() == field)
+    }
+
+    /// Get the `FastFieldWriter` associated with a field.
+    pub fn get_term_id_writer_mut(
+        &mut self,
+        field: Field,
+    ) -> Option<&mut MultiValuedFastFieldWriter> {
+        // TODO optimize
+        self.term_id_writers
+            .iter_mut()
+            .find(|field_writer| field_writer.field() == field)
+    }
+
+    /// Returns the fast field multi-value writer for the given field.
+    ///
+    /// Returns `None` if the field does not exist, or is not
+    /// configured as a multivalued fastfield in the schema.
+    pub fn get_multivalue_writer_mut(
+        &mut self,
+        field: Field,
+    ) -> Option<&mut MultiValuedFastFieldWriter> {
+        // TODO optimize
+        self.multi_values_writers
+            .iter_mut()
+            .find(|multivalue_writer| multivalue_writer.field() == field)
+    }
+
+    /// Returns the bytes fast field writer for the given field.
+    ///
+    /// Returns `None` if the field does not exist, or is not
+    /// configured as a bytes fastfield in the schema.
+    pub fn get_bytes_writer_mut(&mut self, field: Field) -> Option<&mut BytesFastFieldWriter> {
+        // TODO optimize
+        self.bytes_value_writers
+            .iter_mut()
+            .find(|field_writer| field_writer.field() == field)
+    }
     /// Indexes all of the fastfields of a new document.
     pub fn add_document(&mut self, doc: &Document) -> crate::Result<()> {
-        let doc_id = self.num_docs;
-        for field_value in doc.field_values() {
-            if let Some(field_name) =
-                self.fast_field_names[field_value.field().field_id() as usize].as_ref()
-            {
-                match &field_value.value {
-                    Value::U64(u64_val) => {
-                        self.columnar_writer.record_numerical(
-                            doc_id,
-                            field_name.as_str(),
-                            NumericalValue::from(*u64_val),
-                        );
-                    }
-                    Value::I64(i64_val) => {
-                        self.columnar_writer.record_numerical(
-                            doc_id,
-                            field_name.as_str(),
-                            NumericalValue::from(*i64_val),
-                        );
-                    }
-                    Value::F64(f64_val) => {
-                        self.columnar_writer.record_numerical(
-                            doc_id,
-                            field_name.as_str(),
-                            NumericalValue::from(*f64_val),
-                        );
-                    }
-                    Value::Str(_) => todo!(),
-                    Value::PreTokStr(_) => todo!(),
-                    Value::Bool(bool_val) => {
-                        self.columnar_writer
-                            .record_bool(doc_id, field_name.as_str(), *bool_val);
-                    }
-                    Value::Date(datetime) => {
-                        let date_precision =
-                            self.date_precisions[field_value.field().field_id() as usize];
-                        let truncated_datetime = datetime.truncate(date_precision);
-                        self.columnar_writer.record_datetime(
-                            doc_id,
-                            field_name.as_str(),
-                            truncated_datetime.into(),
-                        );
-                    }
-                    Value::Facet(_) => todo!(),
-                    Value::Bytes(_) => todo!(),
-                    Value::JsonObject(_) => todo!(),
-                    Value::IpAddr(_) => todo!(),
-                }
-            }
+        for field_writer in &mut self.term_id_writers {
+            field_writer.add_document(doc)?;
+        }
+        for field_writer in &mut self.single_value_writers {
+            field_writer.add_document(doc)?;
+        }
+        for field_writer in &mut self.multi_values_writers {
+            field_writer.add_document(doc)?;
+        }
+        for field_writer in &mut self.bytes_value_writers {
+            field_writer.add_document(doc)?;
+        }
+        for field_writer in &mut self.u128_value_writers {
+            field_writer.add_document(doc)?;
+        }
+        for field_writer in &mut self.u128_multi_value_writers {
+            field_writer.add_document(doc)?;
         }
-        self.num_docs += 1;
         Ok(())
     }
 
     /// Serializes all of the `FastFieldWriter`s by pushing them in
     /// order to the fast field serializer.
     pub fn serialize(
-        mut self,
-        wrt: &mut dyn io::Write,
+        self,
+        serializer: &mut CompositeFastFieldSerializer,
+        mapping: &HashMap<Field, FxHashMap<UnorderedTermId, TermOrdinal>>,
         doc_id_map: Option<&DocIdMapping>,
     ) -> io::Result<()> {
-        assert!(doc_id_map.is_none()); // TODO handle doc id map
-        let num_docs = self.num_docs;
-        self.columnar_writer.serialize(num_docs, wrt)?;
+        for field_writer in self.term_id_writers {
+            let field = field_writer.field();
+            field_writer.serialize(serializer, mapping.get(&field), doc_id_map)?;
+        }
+        for field_writer in &self.single_value_writers {
+            field_writer.serialize(serializer, doc_id_map)?;
+        }
+
+        for field_writer in self.multi_values_writers {
+            let field = field_writer.field();
+            field_writer.serialize(serializer, mapping.get(&field), doc_id_map)?;
+        }
+        for field_writer in self.bytes_value_writers {
+            field_writer.serialize(serializer, doc_id_map)?;
+        }
+        for field_writer in self.u128_value_writers {
+            field_writer.serialize(serializer, doc_id_map)?;
+        }
+        for field_writer in self.u128_multi_value_writers {
+            field_writer.serialize(serializer, doc_id_map)?;
+        }
+
         Ok(())
     }
 }
+
+/// Fast field writer for u128 values.
+/// The fast field writer just keeps the values in memory.
+///
+/// Only when the segment writer can be closed and
+/// persisted on disk, the fast field writer is
+/// sent to a `FastFieldSerializer` via the `.serialize(...)`
+/// method.
+///
+/// We cannot serialize earlier as the values are
+/// compressed to a compact number space and the number of
+/// bits required for bitpacking can only been known once
+/// we have seen all of the values.
+pub struct U128FastFieldWriter {
+    field: Field,
+    vals: Vec<u128>,
+    val_count: u32,
+}
+
+impl U128FastFieldWriter {
+    /// Creates a new `IntFastFieldWriter`
+    pub fn new(field: Field) -> Self {
+        Self {
+            field,
+            vals: vec![],
+            val_count: 0,
+        }
+    }
+
+    /// The memory used (inclusive childs)
+    pub fn mem_usage(&self) -> usize {
+        self.vals.len() * 16
+    }
+
+    /// Records a new value.
+    ///
+    /// The n-th value being recorded is implicitely
+    /// associated to the document with the `DocId` n.
+    /// (Well, `n-1` actually because of 0-indexing)
+    pub fn add_val(&mut self, val: u128) {
+        self.vals.push(val);
+    }
+
+    /// Extract the fast field value from the document
+    /// (or use the default value) and records it.
+    ///
+    /// Extract the value associated to the fast field for
+    /// this document.
+    pub fn add_document(&mut self, doc: &Document) -> crate::Result<()> {
+        match doc.get_first(self.field) {
+            Some(v) => {
+                let ip_addr = v.as_ip_addr().ok_or_else(|| unexpected_value("ip", v))?;
+                let value = ip_addr.to_u128();
+                self.add_val(value);
+            }
+            None => {
+                self.add_val(0); // TODO fix null handling
+            }
+        };
+        self.val_count += 1;
+        Ok(())
+    }
+
+    /// Push the fast fields value to the `FastFieldWriter`.
+    pub fn serialize(
+        &self,
+        serializer: &mut CompositeFastFieldSerializer,
+        doc_id_map: Option<&DocIdMapping>,
+    ) -> io::Result<()> {
+        if let Some(doc_id_map) = doc_id_map {
+            let iter_gen = || {
+                doc_id_map
+                    .iter_old_doc_ids()
+                    .map(|idx| self.vals[idx as usize])
+            };
+
+            serializer.create_u128_fast_field_with_idx(self.field, iter_gen, self.val_count, 0)?;
+        } else {
+            let iter_gen = || self.vals.iter().cloned();
+            serializer.create_u128_fast_field_with_idx(self.field, iter_gen, self.val_count, 0)?;
+        }
+
+        Ok(())
+    }
+}
+
+/// Fast field writer for ints.
+/// The fast field writer just keeps the values in memory.
+///
+/// Only when the segment writer can be closed and
+/// persisted on disk, the fast field writer is
+/// sent to a `FastFieldSerializer` via the `.serialize(...)`
+/// method.
+///
+/// We cannot serialize earlier as the values are
+/// bitpacked and the number of bits required for bitpacking
+/// can only been known once we have seen all of the values.
+///
+/// Both u64, i64 and f64 use the same writer.
+/// i64 and f64 are just remapped to the `0..2^64 - 1`
+/// using `common::i64_to_u64` and `common::f64_to_u64`.
+pub struct IntFastFieldWriter {
+    field: Field,
+    precision_opt: Option<DatePrecision>,
+    vals: BlockedBitpacker,
+    val_count: usize,
+    val_if_missing: u64,
+    val_min: u64,
+    val_max: u64,
+}
+
+impl IntFastFieldWriter {
+    /// Creates a new `IntFastFieldWriter`
+    pub fn new(field: Field, precision_opt: Option<DatePrecision>) -> IntFastFieldWriter {
+        IntFastFieldWriter {
+            field,
+            precision_opt,
+            vals: BlockedBitpacker::new(),
+            val_count: 0,
+            val_if_missing: 0u64,
+            val_min: u64::MAX,
+            val_max: 0,
+        }
+    }
+
+    /// The memory used (inclusive childs)
+    pub fn mem_usage(&self) -> usize {
+        self.vals.mem_usage()
+    }
+
+    /// Returns the field that this writer is targeting.
+    pub fn field(&self) -> Field {
+        self.field
+    }
+
+    /// Sets the default value.
+    ///
+    /// This default value is recorded for documents if
+    /// a document does not have any value.
+    fn set_val_if_missing(&mut self, val_if_missing: u64) {
+        self.val_if_missing = val_if_missing;
+    }
+
+    /// Records a new value.
+    ///
+    /// The n-th value being recorded is implicitly
+    /// associated with the document with the `DocId` n.
+    /// (Well, `n-1` actually because of 0-indexing)
+    pub fn add_val(&mut self, val: u64) {
+        self.vals.add(val);
+
+        if val > self.val_max {
+            self.val_max = val;
+        }
+        if val < self.val_min {
+            self.val_min = val;
+        }
+
+        self.val_count += 1;
+    }
+
+    /// Extract the fast field value from the document
+    /// (or use the default value) and records it.
+    ///
+    ///
+    /// Extract the value associated with the fast field for
+    /// this document.
+    ///
+    /// i64 and f64 are remapped to u64 using the logic
+    /// in `common::i64_to_u64` and `common::f64_to_u64`.
+    ///
+    /// If the value is missing, then the default value is used
+    /// instead.
+    /// If the document has more than one value for the given field,
+    /// only the first one is taken in account.
+    ///
+    /// Values on text fast fields are skipped.
+    pub fn add_document(&mut self, doc: &Document) -> crate::Result<()> {
+        match doc.get_first(self.field) {
+            Some(v) => {
+                let value = match (self.precision_opt, v) {
+                    (Some(precision), Value::Date(date_val)) => {
+                        date_val.truncate(precision).to_u64()
+                    }
+                    _ => super::value_to_u64(v)?,
+                };
+                self.add_val(value);
+            }
+            None => {
+                self.add_val(self.val_if_missing);
+            }
+        };
+        Ok(())
+    }
+
+    /// get iterator over the data
+    pub(crate) fn iter(&self) -> impl Iterator<Item = u64> + '_ {
+        self.vals.iter()
+    }
+
+    /// Push the fast fields value to the `FastFieldWriter`.
+    pub fn serialize(
+        &self,
+        serializer: &mut CompositeFastFieldSerializer,
+        doc_id_map: Option<&DocIdMapping>,
+    ) -> io::Result<()> {
+        let (min, max) = if self.val_min > self.val_max {
+            (0, 0)
+        } else {
+            (self.val_min, self.val_max)
+        };
+
+        let fastfield_accessor = WriterFastFieldAccessProvider {
+            doc_id_map,
+            vals: &self.vals,
+            min_value: min,
+            max_value: max,
+            num_vals: self.val_count as u32,
+        };
+
+        serializer.create_auto_detect_u64_fast_field(self.field, fastfield_accessor)?;
+
+        Ok(())
+    }
+}
+
+#[derive(Clone)]
+struct WriterFastFieldAccessProvider<'map, 'bitp> {
+    doc_id_map: Option<&'map DocIdMapping>,
+    vals: &'bitp BlockedBitpacker,
+    min_value: u64,
+    max_value: u64,
+    num_vals: u32,
+}
+
+impl<'map, 'bitp> Column for WriterFastFieldAccessProvider<'map, 'bitp> {
+    /// Return the value associated with the given doc.
+    ///
+    /// Whenever possible use the Iterator passed to the fastfield creation instead, for performance
+    /// reasons.
+    ///
+    /// # Panics
+    ///
+    /// May panic if `doc` is greater than the index.
+    fn get_val(&self, _doc: u32) -> u64 {
+        unimplemented!()
+    }
+
+    fn iter(&self) -> Box<dyn Iterator<Item = u64> + '_> {
+        if let Some(doc_id_map) = self.doc_id_map {
+            Box::new(
+                doc_id_map
+                    .iter_old_doc_ids()
+                    .map(|doc_id| self.vals.get(doc_id as usize)),
+            )
+        } else {
+            Box::new(self.vals.iter())
+        }
+    }
+
+    fn min_value(&self) -> u64 {
+        self.min_value
+    }
+
+    fn max_value(&self) -> u64 {
+        self.max_value
+    }
+
+    fn num_vals(&self) -> u32 {
+        self.num_vals
+    }
+}

src/indexer/doc_id_mapping.rs

@@ -113,35 +113,34 @@ pub(crate) fn get_doc_id_mapping_from_field(
     sort_by_field: IndexSortByField,
     segment_writer: &SegmentWriter,
 ) -> crate::Result<DocIdMapping> {
-    todo!()
-    // let schema = segment_writer.segment_serializer.segment().schema();
-    // let field_id = expect_field_id_for_sort_field(&schema, &sort_by_field)?; // for now expect
-    // fastfield, but not strictly required let fast_field = segment_writer
-    //     .fast_field_writers
-    //     .get_field_writer(field_id)
-    //     .ok_or_else(|| {
-    //         TantivyError::InvalidArgument(format!(
-    //             "sort index by field is required to be a fast field {:?}",
-    //             sort_by_field.field
-    //         ))
-    //     })?;
+    let schema = segment_writer.segment_serializer.segment().schema();
+    let field_id = expect_field_id_for_sort_field(&schema, &sort_by_field)?; // for now expect fastfield, but not strictly required
+    let fast_field = segment_writer
+        .fast_field_writers
+        .get_field_writer(field_id)
+        .ok_or_else(|| {
+            TantivyError::InvalidArgument(format!(
+                "sort index by field is required to be a fast field {:?}",
+                sort_by_field.field
+            ))
+        })?;
 
-    // // create new doc_id to old doc_id index (used in fast_field_writers)
-    // let mut doc_id_and_data = fast_field
-    //     .iter()
-    //     .enumerate()
-    //     .map(|el| (el.0 as DocId, el.1))
-    //     .collect::<Vec<_>>();
-    // if sort_by_field.order == Order::Desc {
-    //     doc_id_and_data.sort_by_key(|k| Reverse(k.1));
-    // } else {
-    //     doc_id_and_data.sort_by_key(|k| k.1);
-    // }
-    // let new_doc_id_to_old = doc_id_and_data
-    //     .into_iter()
-    //     .map(|el| el.0)
-    //     .collect::<Vec<_>>();
-    // Ok(DocIdMapping::from_new_id_to_old_id(new_doc_id_to_old))
+    // create new doc_id to old doc_id index (used in fast_field_writers)
+    let mut doc_id_and_data = fast_field
+        .iter()
+        .enumerate()
+        .map(|el| (el.0 as DocId, el.1))
+        .collect::<Vec<_>>();
+    if sort_by_field.order == Order::Desc {
+        doc_id_and_data.sort_by_key(|k| Reverse(k.1));
+    } else {
+        doc_id_and_data.sort_by_key(|k| k.1);
+    }
+    let new_doc_id_to_old = doc_id_and_data
+        .into_iter()
+        .map(|el| el.0)
+        .collect::<Vec<_>>();
+    Ok(DocIdMapping::from_new_id_to_old_id(new_doc_id_to_old))
 }
 
 #[cfg(test)]
@@ -160,11 +159,15 @@ mod tests_indexsorting {
 
         let my_text_field = schema_builder.add_text_field("text_field", text_field_options);
         let my_string_field = schema_builder.add_text_field("string_field", STRING | STORED);
-        let my_number =
-            schema_builder.add_u64_field("my_number", NumericOptions::default().set_fast());
+        let my_number = schema_builder.add_u64_field(
+            "my_number",
+            NumericOptions::default().set_fast(Cardinality::SingleValue),
+        );
 
-        let multi_numbers =
-            schema_builder.add_u64_field("multi_numbers", NumericOptions::default().set_fast());
+        let multi_numbers = schema_builder.add_u64_field(
+            "multi_numbers",
+            NumericOptions::default().set_fast(Cardinality::MultiValues),
+        );
 
         let schema = schema_builder.build();
         let mut index_builder = Index::builder().schema(schema);
@@ -438,48 +441,47 @@ mod tests_indexsorting {
         Ok(())
     }
 
-    // #[test]
-    // fn test_sort_index_fast_field() -> crate::Result<()> {
-    //     let index = create_test_index(
-    //         Some(IndexSettings {
-    //             sort_by_field: Some(IndexSortByField {
-    //                 field: "my_number".to_string(),
-    //                 order: Order::Asc,
-    //             }),
-    //             ..Default::default()
-    //         }),
-    //         get_text_options(),
-    //     )?;
-    //     assert_eq!(
-    //         index.settings().sort_by_field.as_ref().unwrap().field,
-    //         "my_number".to_string()
-    //     );
+    #[test]
+    fn test_sort_index_fast_field() -> crate::Result<()> {
+        let index = create_test_index(
+            Some(IndexSettings {
+                sort_by_field: Some(IndexSortByField {
+                    field: "my_number".to_string(),
+                    order: Order::Asc,
+                }),
+                ..Default::default()
+            }),
+            get_text_options(),
+        )?;
+        assert_eq!(
+            index.settings().sort_by_field.as_ref().unwrap().field,
+            "my_number".to_string()
+        );
 
-    //     let searcher = index.reader()?.searcher();
-    //     assert_eq!(searcher.segment_readers().len(), 1);
-    //     let segment_reader = searcher.segment_reader(0);
-    //     let fast_fields = segment_reader.fast_fields();
-    //     let my_number = index.schema().get_field("my_number").unwrap();
+        let searcher = index.reader()?.searcher();
+        assert_eq!(searcher.segment_readers().len(), 1);
+        let segment_reader = searcher.segment_reader(0);
+        let fast_fields = segment_reader.fast_fields();
+        index.schema().get_field("my_number").unwrap();
 
-    //     let fast_field = fast_fields.u64(my_number).unwrap();
-    //     assert_eq!(fast_field.get_val(0), 10u64);
-    //     assert_eq!(fast_field.get_val(1), 20u64);
-    //     assert_eq!(fast_field.get_val(2), 30u64);
+        let fast_field = fast_fields.u64("my_number").unwrap();
+        assert_eq!(fast_field.get_val(0), 10u64);
+        assert_eq!(fast_field.get_val(1), 20u64);
+        assert_eq!(fast_field.get_val(2), 30u64);
 
-    //     let multi_numbers = index.schema().get_field("multi_numbers").unwrap();
-    //     let multifield = fast_fields.u64s(multi_numbers).unwrap();
-    //     let mut vals = vec![];
-    //     multifield.get_vals(0u32, &mut vals);
-    //     assert_eq!(vals, &[] as &[u64]);
-    //     let mut vals = vec![];
-    //     multifield.get_vals(1u32, &mut vals);
-    //     assert_eq!(vals, &[5, 6]);
+        let multifield = fast_fields.u64s("multi_numbers").unwrap();
+        let mut vals = vec![];
+        multifield.get_vals(0u32, &mut vals);
+        assert_eq!(vals, &[] as &[u64]);
+        let mut vals = vec![];
+        multifield.get_vals(1u32, &mut vals);
+        assert_eq!(vals, &[5, 6]);
 
-    //     let mut vals = vec![];
-    //     multifield.get_vals(2u32, &mut vals);
-    //     assert_eq!(vals, &[3]);
-    //     Ok(())
-    // }
+        let mut vals = vec![];
+        multifield.get_vals(2u32, &mut vals);
+        assert_eq!(vals, &[3]);
+        Ok(())
+    }
 
     #[test]
     fn test_doc_mapping() {

src/indexer/json_term_writer.rs

@@ -150,6 +150,7 @@ fn index_json_value(
                     json_term_writer.term_buffer,
                     ctx,
                     indexing_position,
+                    None,
                 );
             }
             TextOrDateTime::DateTime(dt) => {

src/indexer/merger_sorted_index_test.rs

@@ -2,17 +2,19 @@
 mod tests {
     use crate::collector::TopDocs;
     use crate::core::Index;
-    use crate::fastfield::AliveBitSet;
+    use crate::fastfield::{AliveBitSet, MultiValuedFastFieldReader};
     use crate::query::QueryParser;
     use crate::schema::{
-        self, BytesOptions, Facet, FacetOptions, IndexRecordOption, NumericOptions,
+        self, BytesOptions, Cardinality, Facet, FacetOptions, IndexRecordOption, NumericOptions,
         TextFieldIndexing, TextOptions,
     };
     use crate::{DocAddress, DocSet, IndexSettings, IndexSortByField, Order, Postings, Term};
 
     fn create_test_index_posting_list_issue(index_settings: Option<IndexSettings>) -> Index {
         let mut schema_builder = schema::Schema::builder();
-        let int_options = NumericOptions::default().set_fast().set_indexed();
+        let int_options = NumericOptions::default()
+            .set_fast(Cardinality::SingleValue)
+            .set_indexed();
         let int_field = schema_builder.add_u64_field("intval", int_options);
 
         let facet_field = schema_builder.add_facet_field("facet", FacetOptions::default());
@@ -60,7 +62,7 @@ mod tests {
     ) -> crate::Result<Index> {
         let mut schema_builder = schema::Schema::builder();
         let int_options = NumericOptions::default()
-            .set_fast()
+            .set_fast(Cardinality::SingleValue)
             .set_stored()
             .set_indexed();
         let int_field = schema_builder.add_u64_field("intval", int_options);
@@ -69,8 +71,10 @@ mod tests {
         let bytes_field = schema_builder.add_bytes_field("bytes", bytes_options);
         let facet_field = schema_builder.add_facet_field("facet", FacetOptions::default());
 
-        let multi_numbers =
-            schema_builder.add_u64_field("multi_numbers", NumericOptions::default().set_fast());
+        let multi_numbers = schema_builder.add_u64_field(
+            "multi_numbers",
+            NumericOptions::default().set_fast(Cardinality::MultiValues),
+        );
         let text_field_options = TextOptions::default()
             .set_indexing_options(
                 TextFieldIndexing::default()
@@ -345,130 +349,128 @@ mod tests {
         }
     }
 
-    // #[test]
-    // fn test_merge_sorted_index_asc() {
-    //     let index = create_test_index(
-    //         Some(IndexSettings {
-    //             sort_by_field: Some(IndexSortByField {
-    //                 field: "intval".to_string(),
-    //                 order: Order::Asc,
-    //             }),
-    //             ..Default::default()
-    //         }),
-    //         false,
-    //     )
-    //     .unwrap();
+    #[test]
+    fn test_merge_sorted_index_asc() {
+        let index = create_test_index(
+            Some(IndexSettings {
+                sort_by_field: Some(IndexSortByField {
+                    field: "intval".to_string(),
+                    order: Order::Asc,
+                }),
+                ..Default::default()
+            }),
+            false,
+        )
+        .unwrap();
 
-    //     let int_field = index.schema().get_field("intval").unwrap();
-    //     let multi_numbers = index.schema().get_field("multi_numbers").unwrap();
-    //     let bytes_field = index.schema().get_field("bytes").unwrap();
-    //     let reader = index.reader().unwrap();
-    //     let searcher = reader.searcher();
-    //     assert_eq!(searcher.segment_readers().len(), 1);
-    //     let segment_reader = searcher.segment_readers().last().unwrap();
+        let int_field = index.schema().get_field("intval").unwrap();
+        let reader = index.reader().unwrap();
+        let searcher = reader.searcher();
+        assert_eq!(searcher.segment_readers().len(), 1);
+        let segment_reader = searcher.segment_readers().last().unwrap();
 
-    //     let fast_fields = segment_reader.fast_fields();
-    //     let fast_field = fast_fields.u64(int_field).unwrap();
-    //     assert_eq!(fast_field.get_val(0), 1u64);
-    //     assert_eq!(fast_field.get_val(1), 2u64);
-    //     assert_eq!(fast_field.get_val(2), 3u64);
-    //     assert_eq!(fast_field.get_val(3), 10u64);
-    //     assert_eq!(fast_field.get_val(4), 20u64);
-    //     assert_eq!(fast_field.get_val(5), 1_000u64);
+        let fast_fields = segment_reader.fast_fields();
+        let fast_field = fast_fields.u64("intval").unwrap();
+        assert_eq!(fast_field.get_val(0), 1u64);
+        assert_eq!(fast_field.get_val(1), 2u64);
+        assert_eq!(fast_field.get_val(2), 3u64);
+        assert_eq!(fast_field.get_val(3), 10u64);
+        assert_eq!(fast_field.get_val(4), 20u64);
+        assert_eq!(fast_field.get_val(5), 1_000u64);
 
-    //     let get_vals = |fast_field: &MultiValuedFastFieldReader<u64>, doc_id: u32| -> Vec<u64> {
-    //         let mut vals = vec![];
-    //         fast_field.get_vals(doc_id, &mut vals);
-    //         vals
-    //     };
-    //     let fast_fields = segment_reader.fast_fields();
-    //     let fast_field = fast_fields.u64s(multi_numbers).unwrap();
-    //     assert_eq!(&get_vals(&fast_field, 0), &[] as &[u64]);
-    //     assert_eq!(&get_vals(&fast_field, 1), &[2, 3]);
-    //     assert_eq!(&get_vals(&fast_field, 2), &[3, 4]);
-    //     assert_eq!(&get_vals(&fast_field, 3), &[10, 11]);
-    //     assert_eq!(&get_vals(&fast_field, 4), &[20]);
-    //     assert_eq!(&get_vals(&fast_field, 5), &[1001, 1002]);
+        let get_vals = |fast_field: &MultiValuedFastFieldReader<u64>, doc_id: u32| -> Vec<u64> {
+            let mut vals = vec![];
+            fast_field.get_vals(doc_id, &mut vals);
+            vals
+        };
+        let fast_fields = segment_reader.fast_fields();
+        let fast_field = fast_fields.u64s("multi_numbers").unwrap();
+        assert_eq!(&get_vals(&fast_field, 0), &[] as &[u64]);
+        assert_eq!(&get_vals(&fast_field, 1), &[2, 3]);
+        assert_eq!(&get_vals(&fast_field, 2), &[3, 4]);
+        assert_eq!(&get_vals(&fast_field, 3), &[10, 11]);
+        assert_eq!(&get_vals(&fast_field, 4), &[20]);
+        assert_eq!(&get_vals(&fast_field, 5), &[1001, 1002]);
 
-    //     let fast_field = fast_fields.bytes(bytes_field).unwrap();
-    //     assert_eq!(fast_field.get_bytes(0), &[] as &[u8]);
-    //     assert_eq!(fast_field.get_bytes(2), &[1, 2, 3]);
-    //     assert_eq!(fast_field.get_bytes(5), &[5, 5]);
+        let fast_field = fast_fields.bytes("bytes").unwrap();
+        assert_eq!(fast_field.get_bytes(0), &[] as &[u8]);
+        assert_eq!(fast_field.get_bytes(2), &[1, 2, 3]);
+        assert_eq!(fast_field.get_bytes(5), &[5, 5]);
 
-    //     // test new field norm mapping
-    //     {
-    //         let my_text_field = index.schema().get_field("text_field").unwrap();
-    //         let fieldnorm_reader = segment_reader.get_fieldnorms_reader(my_text_field).unwrap();
-    //         assert_eq!(fieldnorm_reader.fieldnorm(0), 0);
-    //         assert_eq!(fieldnorm_reader.fieldnorm(1), 4);
-    //         assert_eq!(fieldnorm_reader.fieldnorm(2), 2); // some text
-    //         assert_eq!(fieldnorm_reader.fieldnorm(3), 1);
-    //         assert_eq!(fieldnorm_reader.fieldnorm(5), 3); // the biggest num
-    //     }
+        // test new field norm mapping
+        {
+            let my_text_field = index.schema().get_field("text_field").unwrap();
+            let fieldnorm_reader = segment_reader.get_fieldnorms_reader(my_text_field).unwrap();
+            assert_eq!(fieldnorm_reader.fieldnorm(0), 0);
+            assert_eq!(fieldnorm_reader.fieldnorm(1), 4);
+            assert_eq!(fieldnorm_reader.fieldnorm(2), 2); // some text
+            assert_eq!(fieldnorm_reader.fieldnorm(3), 1);
+            assert_eq!(fieldnorm_reader.fieldnorm(5), 3); // the biggest num
+        }
 
-    //     let searcher = index.reader().unwrap().searcher();
-    //     {
-    //         let my_text_field = index.schema().get_field("text_field").unwrap();
+        let searcher = index.reader().unwrap().searcher();
+        {
+            let my_text_field = index.schema().get_field("text_field").unwrap();
 
-    //         let do_search = |term: &str| {
-    //             let query = QueryParser::for_index(&index, vec![my_text_field])
-    //                 .parse_query(term)
-    //                 .unwrap();
-    //             let top_docs: Vec<(f32, DocAddress)> =
-    //                 searcher.search(&query, &TopDocs::with_limit(3)).unwrap();
+            let do_search = |term: &str| {
+                let query = QueryParser::for_index(&index, vec![my_text_field])
+                    .parse_query(term)
+                    .unwrap();
+                let top_docs: Vec<(f32, DocAddress)> =
+                    searcher.search(&query, &TopDocs::with_limit(3)).unwrap();
 
-    //             top_docs.iter().map(|el| el.1.doc_id).collect::<Vec<_>>()
-    //         };
+                top_docs.iter().map(|el| el.1.doc_id).collect::<Vec<_>>()
+            };
 
-    //         assert_eq!(do_search("some"), vec![2]);
-    //         assert_eq!(do_search("blubber"), vec![3]);
-    //         assert_eq!(do_search("biggest"), vec![5]);
-    //     }
+            assert_eq!(do_search("some"), vec![2]);
+            assert_eq!(do_search("blubber"), vec![3]);
+            assert_eq!(do_search("biggest"), vec![5]);
+        }
 
-    //     // postings file
-    //     {
-    //         let my_text_field = index.schema().get_field("text_field").unwrap();
-    //         let term_a = Term::from_field_text(my_text_field, "text");
-    //         let inverted_index = segment_reader.inverted_index(my_text_field).unwrap();
-    //         let mut postings = inverted_index
-    //             .read_postings(&term_a, IndexRecordOption::WithFreqsAndPositions)
-    //             .unwrap()
-    //             .unwrap();
+        // postings file
+        {
+            let my_text_field = index.schema().get_field("text_field").unwrap();
+            let term_a = Term::from_field_text(my_text_field, "text");
+            let inverted_index = segment_reader.inverted_index(my_text_field).unwrap();
+            let mut postings = inverted_index
+                .read_postings(&term_a, IndexRecordOption::WithFreqsAndPositions)
+                .unwrap()
+                .unwrap();
 
-    //         assert_eq!(postings.doc_freq(), 2);
-    //         let fallback_bitset = AliveBitSet::for_test_from_deleted_docs(&[0], 100);
-    //         assert_eq!(
-    //             postings.doc_freq_given_deletes(
-    //                 segment_reader.alive_bitset().unwrap_or(&fallback_bitset)
-    //             ),
-    //             2
-    //         );
+            assert_eq!(postings.doc_freq(), 2);
+            let fallback_bitset = AliveBitSet::for_test_from_deleted_docs(&[0], 100);
+            assert_eq!(
+                postings.doc_freq_given_deletes(
+                    segment_reader.alive_bitset().unwrap_or(&fallback_bitset)
+                ),
+                2
+            );
 
-    //         let mut output = vec![];
-    //         postings.positions(&mut output);
-    //         assert_eq!(output, vec![1, 3]);
-    //         postings.advance();
+            let mut output = vec![];
+            postings.positions(&mut output);
+            assert_eq!(output, vec![1, 3]);
+            postings.advance();
 
-    //         postings.positions(&mut output);
-    //         assert_eq!(output, vec![1]);
-    //     }
+            postings.positions(&mut output);
+            assert_eq!(output, vec![1]);
+        }
 
-    //     // access doc store
-    //     {
-    //         let doc = searcher.doc(DocAddress::new(0, 0)).unwrap();
-    //         assert_eq!(doc.get_first(int_field).unwrap().as_u64(), Some(1));
-    //         let doc = searcher.doc(DocAddress::new(0, 1)).unwrap();
-    //         assert_eq!(doc.get_first(int_field).unwrap().as_u64(), Some(2));
-    //         let doc = searcher.doc(DocAddress::new(0, 2)).unwrap();
-    //         assert_eq!(doc.get_first(int_field).unwrap().as_u64(), Some(3));
-    //         let doc = searcher.doc(DocAddress::new(0, 3)).unwrap();
-    //         assert_eq!(doc.get_first(int_field).unwrap().as_u64(), Some(10));
-    //         let doc = searcher.doc(DocAddress::new(0, 4)).unwrap();
-    //         assert_eq!(doc.get_first(int_field).unwrap().as_u64(), Some(20));
-    //         let doc = searcher.doc(DocAddress::new(0, 5)).unwrap();
-    //         assert_eq!(doc.get_first(int_field).unwrap().as_u64(), Some(1_000));
-    //     }
-    // }
+        // access doc store
+        {
+            let doc = searcher.doc(DocAddress::new(0, 0)).unwrap();
+            assert_eq!(doc.get_first(int_field).unwrap().as_u64(), Some(1));
+            let doc = searcher.doc(DocAddress::new(0, 1)).unwrap();
+            assert_eq!(doc.get_first(int_field).unwrap().as_u64(), Some(2));
+            let doc = searcher.doc(DocAddress::new(0, 2)).unwrap();
+            assert_eq!(doc.get_first(int_field).unwrap().as_u64(), Some(3));
+            let doc = searcher.doc(DocAddress::new(0, 3)).unwrap();
+            assert_eq!(doc.get_first(int_field).unwrap().as_u64(), Some(10));
+            let doc = searcher.doc(DocAddress::new(0, 4)).unwrap();
+            assert_eq!(doc.get_first(int_field).unwrap().as_u64(), Some(20));
+            let doc = searcher.doc(DocAddress::new(0, 5)).unwrap();
+            assert_eq!(doc.get_first(int_field).unwrap().as_u64(), Some(1_000));
+        }
+    }
 }
 
 #[cfg(all(test, feature = "unstable"))]
@@ -485,7 +487,9 @@ mod bench_sorted_index_merge {
     use crate::{IndexSettings, IndexSortByField, IndexWriter, Order};
     fn create_index(sort_by_field: Option<IndexSortByField>) -> Index {
         let mut schema_builder = Schema::builder();
-        let int_options = NumericOptions::default().set_fast().set_indexed();
+        let int_options = NumericOptions::default()
+            .set_fast(Cardinality::SingleValue)
+            .set_indexed();
         let int_field = schema_builder.add_u64_field("intval", int_options);
         let schema = schema_builder.build();
 

src/indexer/mod.rs

@@ -19,8 +19,8 @@ mod segment_register;
 pub mod segment_serializer;
 pub mod segment_updater;
 mod segment_writer;
-// mod sorted_doc_id_column;
-// mod sorted_doc_id_multivalue_column;
+mod sorted_doc_id_column;
+mod sorted_doc_id_multivalue_column;
 mod stamper;
 
 use crossbeam_channel as channel;
@@ -58,7 +58,7 @@ type AddBatchReceiver = channel::Receiver<AddBatch>;
 #[cfg(test)]
 mod tests_mmap {
     use crate::collector::Count;
-    // use crate::query::QueryParser;
+    use crate::query::QueryParser;
     use crate::schema::{JsonObjectOptions, Schema, TEXT};
     use crate::{Index, Term};
 
@@ -79,45 +79,45 @@ mod tests_mmap {
         Ok(())
     }
 
-    // #[test]
-    // fn test_json_field_expand_dots_disabled_dot_escaped_required() {
-    //     let mut schema_builder = Schema::builder();
-    //     let json_field = schema_builder.add_json_field("json", TEXT);
-    //     let index = Index::create_in_ram(schema_builder.build());
-    //     let mut index_writer = index.writer_for_tests().unwrap();
-    //     let json = serde_json::json!({"k8s.container.name": "prometheus", "val": "hello"});
-    //     index_writer.add_document(doc!(json_field=>json)).unwrap();
-    //     index_writer.commit().unwrap();
-    //     let reader = index.reader().unwrap();
-    //     let searcher = reader.searcher();
-    //     assert_eq!(searcher.num_docs(), 1);
-    //     let parse_query = QueryParser::for_index(&index, Vec::new());
-    //     let query = parse_query
-    //         .parse_query(r#"json.k8s\.container\.name:prometheus"#)
-    //         .unwrap();
-    //     let num_docs = searcher.search(&query, &Count).unwrap();
-    //     assert_eq!(num_docs, 1);
-    // }
+    #[test]
+    fn test_json_field_expand_dots_disabled_dot_escaped_required() {
+        let mut schema_builder = Schema::builder();
+        let json_field = schema_builder.add_json_field("json", TEXT);
+        let index = Index::create_in_ram(schema_builder.build());
+        let mut index_writer = index.writer_for_tests().unwrap();
+        let json = serde_json::json!({"k8s.container.name": "prometheus", "val": "hello"});
+        index_writer.add_document(doc!(json_field=>json)).unwrap();
+        index_writer.commit().unwrap();
+        let reader = index.reader().unwrap();
+        let searcher = reader.searcher();
+        assert_eq!(searcher.num_docs(), 1);
+        let parse_query = QueryParser::for_index(&index, Vec::new());
+        let query = parse_query
+            .parse_query(r#"json.k8s\.container\.name:prometheus"#)
+            .unwrap();
+        let num_docs = searcher.search(&query, &Count).unwrap();
+        assert_eq!(num_docs, 1);
+    }
 
-    // #[test]
-    // fn test_json_field_expand_dots_enabled_dot_escape_not_required() {
-    //     let mut schema_builder = Schema::builder();
-    //     let json_options: JsonObjectOptions =
-    //         JsonObjectOptions::from(TEXT).set_expand_dots_enabled();
-    //     let json_field = schema_builder.add_json_field("json", json_options);
-    //     let index = Index::create_in_ram(schema_builder.build());
-    //     let mut index_writer = index.writer_for_tests().unwrap();
-    //     let json = serde_json::json!({"k8s.container.name": "prometheus", "val": "hello"});
-    //     index_writer.add_document(doc!(json_field=>json)).unwrap();
-    //     index_writer.commit().unwrap();
-    //     let reader = index.reader().unwrap();
-    //     let searcher = reader.searcher();
-    //     assert_eq!(searcher.num_docs(), 1);
-    //     let parse_query = QueryParser::for_index(&index, Vec::new());
-    //     let query = parse_query
-    //         .parse_query(r#"json.k8s.container.name:prometheus"#)
-    //         .unwrap();
-    //     let num_docs = searcher.search(&query, &Count).unwrap();
-    //     assert_eq!(num_docs, 1);
-    // }
+    #[test]
+    fn test_json_field_expand_dots_enabled_dot_escape_not_required() {
+        let mut schema_builder = Schema::builder();
+        let json_options: JsonObjectOptions =
+            JsonObjectOptions::from(TEXT).set_expand_dots_enabled();
+        let json_field = schema_builder.add_json_field("json", json_options);
+        let index = Index::create_in_ram(schema_builder.build());
+        let mut index_writer = index.writer_for_tests().unwrap();
+        let json = serde_json::json!({"k8s.container.name": "prometheus", "val": "hello"});
+        index_writer.add_document(doc!(json_field=>json)).unwrap();
+        index_writer.commit().unwrap();
+        let reader = index.reader().unwrap();
+        let searcher = reader.searcher();
+        assert_eq!(searcher.num_docs(), 1);
+        let parse_query = QueryParser::for_index(&index, Vec::new());
+        let query = parse_query
+            .parse_query(r#"json.k8s.container.name:prometheus"#)
+            .unwrap();
+        let num_docs = searcher.search(&query, &Count).unwrap();
+        assert_eq!(num_docs, 1);
+    }
 }

src/indexer/segment_serializer.rs

@@ -1,7 +1,4 @@
-use common::TerminatingWrite;
-
 use crate::core::{Segment, SegmentComponent};
-use crate::directory::WritePtr;
 use crate::fastfield::CompositeFastFieldSerializer;
 use crate::fieldnorm::FieldNormsSerializer;
 use crate::postings::InvertedIndexSerializer;
@@ -12,7 +9,7 @@ use crate::store::StoreWriter;
 pub struct SegmentSerializer {
     segment: Segment,
     pub(crate) store_writer: StoreWriter,
-    fast_field_write: WritePtr,
+    fast_field_serializer: CompositeFastFieldSerializer,
     fieldnorms_serializer: Option<FieldNormsSerializer>,
     postings_serializer: InvertedIndexSerializer,
 }
@@ -50,6 +47,7 @@ impl SegmentSerializer {
         };
 
         let fast_field_write = segment.open_write(SegmentComponent::FastFields)?;
+        let fast_field_serializer = CompositeFastFieldSerializer::from_write(fast_field_write)?;
 
         let fieldnorms_write = segment.open_write(SegmentComponent::FieldNorms)?;
         let fieldnorms_serializer = FieldNormsSerializer::from_write(fieldnorms_write)?;
@@ -58,7 +56,7 @@ impl SegmentSerializer {
         Ok(SegmentSerializer {
             segment,
             store_writer,
-            fast_field_write,
+            fast_field_serializer,
             fieldnorms_serializer: Some(fieldnorms_serializer),
             postings_serializer,
         })
@@ -83,8 +81,8 @@ impl SegmentSerializer {
     }
 
     /// Accessor to the `FastFieldSerializer`.
-    pub fn get_fast_field_write(&mut self) -> &mut WritePtr {
-        &mut self.fast_field_write
+    pub fn get_fast_field_serializer(&mut self) -> &mut CompositeFastFieldSerializer {
+        &mut self.fast_field_serializer
     }
 
     /// Extract the field norm serializer.
@@ -104,7 +102,7 @@ impl SegmentSerializer {
         if let Some(fieldnorms_serializer) = self.extract_fieldnorms_serializer() {
             fieldnorms_serializer.close()?;
         }
-        self.fast_field_write.terminate()?;
+        self.fast_field_serializer.close()?;
         self.postings_serializer.close()?;
         self.store_writer.close()?;
         Ok(())

src/indexer/segment_writer.rs

@@ -139,6 +139,7 @@ impl SegmentWriter {
             self.ctx,
             self.fast_field_writers,
             &self.fieldnorms_writer,
+            &self.schema,
             self.segment_serializer,
             mapping.as_ref(),
         )?;
@@ -184,15 +185,22 @@ impl SegmentWriter {
                     for value in values {
                         let facet = value.as_facet().ok_or_else(make_schema_error)?;
                         let facet_str = facet.encoded_str();
-                        let mut facet_tokenizer = FacetTokenizer.token_stream(facet_str);
-                        let mut indexing_position = IndexingPosition::default();
-                        postings_writer.index_text(
-                            doc_id,
-                            &mut *facet_tokenizer,
-                            term_buffer,
-                            ctx,
-                            &mut indexing_position,
-                        );
+                        let mut unordered_term_id_opt = None;
+                        FacetTokenizer
+                            .token_stream(facet_str)
+                            .process(&mut |token| {
+                                term_buffer.set_text(&token.text);
+                                let unordered_term_id =
+                                    postings_writer.subscribe(doc_id, 0u32, term_buffer, ctx);
+                                // TODO pass indexing context directly in subscribe function
+                                unordered_term_id_opt = Some(unordered_term_id);
+                            });
+                        if let Some(unordered_term_id) = unordered_term_id_opt {
+                            self.fast_field_writers
+                                .get_term_id_writer_mut(field)
+                                .expect("writer for facet missing")
+                                .add_val(unordered_term_id);
+                        }
                     }
                 }
                 FieldType::Str(_) => {
@@ -219,6 +227,7 @@ impl SegmentWriter {
                             term_buffer,
                             ctx,
                             &mut indexing_position,
+                            self.fast_field_writers.get_term_id_writer_mut(field),
                         );
                     }
                     if field_entry.has_fieldnorms() {
@@ -374,6 +383,7 @@ fn remap_and_write(
     ctx: IndexingContext,
     fast_field_writers: FastFieldsWriter,
     fieldnorms_writer: &FieldNormsWriter,
+    schema: &Schema,
     mut serializer: SegmentSerializer,
     doc_id_map: Option<&DocIdMapping>,
 ) -> crate::Result<()> {
@@ -385,15 +395,20 @@ fn remap_and_write(
         .segment()
         .open_read(SegmentComponent::FieldNorms)?;
     let fieldnorm_readers = FieldNormReaders::open(fieldnorm_data)?;
-    serialize_postings(
+    let term_ord_map = serialize_postings(
         ctx,
         per_field_postings_writers,
         fieldnorm_readers,
         doc_id_map,
+        schema,
         serializer.get_postings_serializer(),
     )?;
     debug!("fastfield-serialize");
-    fast_field_writers.serialize(serializer.get_fast_field_write(), doc_id_map)?;
+    fast_field_writers.serialize(
+        serializer.get_fast_field_serializer(),
+        &term_ord_map,
+        doc_id_map,
+    )?;
 
     // finalize temp docstore and create version, which reflects the doc_id_map
     if let Some(doc_id_map) = doc_id_map {

src/lib.rs

@@ -147,22 +147,6 @@ pub struct DateTime {
     pub(crate) timestamp_micros: i64,
 }
 
-impl From<columnar::DateTime> for DateTime {
-    fn from(columnar_datetime: columnar::DateTime) -> Self {
-        DateTime {
-            timestamp_micros: columnar_datetime.timestamp_micros,
-        }
-    }
-}
-
-impl From<DateTime> for columnar::DateTime {
-    fn from(datetime: crate::DateTime) -> Self {
-        columnar::DateTime {
-            timestamp_micros: datetime.timestamp_micros,
-        }
-    }
-}
-
 impl DateTime {
     /// Create new from UNIX timestamp in seconds
     pub const fn from_timestamp_secs(seconds: i64) -> Self {
@@ -279,7 +263,7 @@ mod indexer;
 pub mod error;
 pub mod tokenizer;
 
-// pub mod aggregation;
+pub mod aggregation;
 pub mod collector;
 pub mod directory;
 pub mod fastfield;

src/postings/json_postings_writer.rs

@@ -2,10 +2,13 @@ use std::io;
 
 use stacker::Addr;
 
+use crate::fastfield::MultiValuedFastFieldWriter;
 use crate::indexer::doc_id_mapping::DocIdMapping;
 use crate::postings::postings_writer::SpecializedPostingsWriter;
 use crate::postings::recorder::{BufferLender, DocIdRecorder, Recorder};
-use crate::postings::{FieldSerializer, IndexingContext, IndexingPosition, PostingsWriter};
+use crate::postings::{
+    FieldSerializer, IndexingContext, IndexingPosition, PostingsWriter, UnorderedTermId,
+};
 use crate::schema::term::as_json_path_type_value_bytes;
 use crate::schema::Type;
 use crate::tokenizer::TokenStream;
@@ -30,8 +33,8 @@ impl<Rec: Recorder> PostingsWriter for JsonPostingsWriter<Rec> {
         pos: u32,
         term: &crate::Term,
         ctx: &mut IndexingContext,
-    ) {
-        self.non_str_posting_writer.subscribe(doc, pos, term, ctx);
+    ) -> UnorderedTermId {
+        self.non_str_posting_writer.subscribe(doc, pos, term, ctx)
     }
 
     fn index_text(
@@ -41,6 +44,7 @@ impl<Rec: Recorder> PostingsWriter for JsonPostingsWriter<Rec> {
         term_buffer: &mut Term,
         ctx: &mut IndexingContext,
         indexing_position: &mut IndexingPosition,
+        _fast_field_writer: Option<&mut MultiValuedFastFieldWriter>,
     ) {
         self.str_posting_writer.index_text(
             doc_id,
@@ -48,19 +52,20 @@ impl<Rec: Recorder> PostingsWriter for JsonPostingsWriter<Rec> {
             term_buffer,
             ctx,
             indexing_position,
+            None,
         );
     }
 
     /// The actual serialization format is handled by the `PostingsSerializer`.
     fn serialize(
         &self,
-        term_addrs: &[(Term<&[u8]>, Addr)],
+        term_addrs: &[(Term<&[u8]>, Addr, UnorderedTermId)],
         doc_id_map: Option<&DocIdMapping>,
         ctx: &IndexingContext,
         serializer: &mut FieldSerializer,
     ) -> io::Result<()> {
         let mut buffer_lender = BufferLender::default();
-        for (term, addr) in term_addrs {
+        for (term, addr, _) in term_addrs {
             // TODO optimization opportunity here.
             if let Some((_, typ, _)) = as_json_path_type_value_bytes(term.value_bytes()) {
                 if typ == Type::Str {

src/postings/postings_writer.rs

@@ -6,6 +6,7 @@ use std::ops::Range;
 use rustc_hash::FxHashMap;
 use stacker::Addr;
 
+use crate::fastfield::MultiValuedFastFieldWriter;
 use crate::fieldnorm::FieldNormReaders;
 use crate::indexer::doc_id_mapping::DocIdMapping;
 use crate::postings::recorder::{BufferLender, Recorder};
@@ -20,10 +21,12 @@ use crate::DocId;
 
 const POSITION_GAP: u32 = 1;
 
-fn make_field_partition(term_offsets: &[(Term<&[u8]>, Addr)]) -> Vec<(Field, Range<usize>)> {
+fn make_field_partition(
+    term_offsets: &[(Term<&[u8]>, Addr, UnorderedTermId)],
+) -> Vec<(Field, Range<usize>)> {
     let term_offsets_it = term_offsets
         .iter()
-        .map(|(term, _)| term.field())
+        .map(|(term, _, _)| term.field())
         .enumerate();
     let mut prev_field_opt = None;
     let mut fields = vec![];
@@ -51,18 +54,48 @@ pub(crate) fn serialize_postings(
     per_field_postings_writers: &PerFieldPostingsWriter,
     fieldnorm_readers: FieldNormReaders,
     doc_id_map: Option<&DocIdMapping>,
+    schema: &Schema,
     serializer: &mut InvertedIndexSerializer,
-) -> crate::Result<()> {
-    let mut term_offsets: Vec<(Term<&[u8]>, Addr)> = Vec::with_capacity(ctx.term_index.len());
+) -> crate::Result<HashMap<Field, FxHashMap<UnorderedTermId, TermOrdinal>>> {
+    let mut term_offsets: Vec<(Term<&[u8]>, Addr, UnorderedTermId)> =
+        Vec::with_capacity(ctx.term_index.len());
     term_offsets.extend(
         ctx.term_index
             .iter()
-            .map(|(bytes, addr, _unordered_id)| (Term::wrap(bytes), addr)),
+            .map(|(bytes, addr, unordered_id)| (Term::wrap(bytes), addr, unordered_id)),
     );
-    term_offsets.sort_unstable_by_key(|(k, _)| k.clone());
+    term_offsets.sort_unstable_by_key(|(k, _, _)| k.clone());
+    let mut unordered_term_mappings: HashMap<Field, FxHashMap<UnorderedTermId, TermOrdinal>> =
+        HashMap::new();
 
     let field_offsets = make_field_partition(&term_offsets);
     for (field, byte_offsets) in field_offsets {
+        let field_entry = schema.get_field_entry(field);
+        match *field_entry.field_type() {
+            FieldType::Str(_) | FieldType::Facet(_) => {
+                // populating the (unordered term ord) -> (ordered term ord) mapping
+                // for the field.
+                let unordered_term_ids = term_offsets[byte_offsets.clone()]
+                    .iter()
+                    .map(|&(_, _, bucket)| bucket);
+                let mapping: FxHashMap<UnorderedTermId, TermOrdinal> = unordered_term_ids
+                    .enumerate()
+                    .map(|(term_ord, unord_term_id)| {
+                        (unord_term_id as UnorderedTermId, term_ord as TermOrdinal)
+                    })
+                    .collect();
+                unordered_term_mappings.insert(field, mapping);
+            }
+            FieldType::U64(_)
+            | FieldType::I64(_)
+            | FieldType::F64(_)
+            | FieldType::Date(_)
+            | FieldType::Bool(_) => {}
+            FieldType::Bytes(_) => {}
+            FieldType::JsonObject(_) => {}
+            FieldType::IpAddr(_) => {}
+        }
+
         let postings_writer = per_field_postings_writers.get_for_field(field);
         let fieldnorm_reader = fieldnorm_readers.get_field(field)?;
         let mut field_serializer =
@@ -75,7 +108,7 @@ pub(crate) fn serialize_postings(
         )?;
         field_serializer.close()?;
     }
-    Ok(())
+    Ok(unordered_term_mappings)
 }
 
 #[derive(Default)]
@@ -96,13 +129,19 @@ pub(crate) trait PostingsWriter: Send + Sync {
     /// * term - the term
     /// * ctx - Contains a term hashmap and a memory arena to store all necessary posting list
     ///   information.
-    fn subscribe(&mut self, doc: DocId, pos: u32, term: &Term, ctx: &mut IndexingContext);
+    fn subscribe(
+        &mut self,
+        doc: DocId,
+        pos: u32,
+        term: &Term,
+        ctx: &mut IndexingContext,
+    ) -> UnorderedTermId;
 
     /// Serializes the postings on disk.
     /// The actual serialization format is handled by the `PostingsSerializer`.
     fn serialize(
         &self,
-        term_addrs: &[(Term<&[u8]>, Addr)],
+        term_addrs: &[(Term<&[u8]>, Addr, UnorderedTermId)],
         doc_id_map: Option<&DocIdMapping>,
         ctx: &IndexingContext,
         serializer: &mut FieldSerializer,
@@ -116,6 +155,7 @@ pub(crate) trait PostingsWriter: Send + Sync {
         term_buffer: &mut Term,
         ctx: &mut IndexingContext,
         indexing_position: &mut IndexingPosition,
+        mut term_id_fast_field_writer_opt: Option<&mut MultiValuedFastFieldWriter>,
     ) {
         let end_of_path_idx = term_buffer.len_bytes();
         let mut num_tokens = 0;
@@ -135,7 +175,11 @@ pub(crate) trait PostingsWriter: Send + Sync {
             term_buffer.append_bytes(token.text.as_bytes());
             let start_position = indexing_position.end_position + token.position as u32;
             end_position = end_position.max(start_position + token.position_length as u32);
-            self.subscribe(doc_id, start_position, term_buffer, ctx);
+            let unordered_term_id = self.subscribe(doc_id, start_position, term_buffer, ctx);
+            if let Some(term_id_fast_field_writer) = term_id_fast_field_writer_opt.as_mut() {
+                term_id_fast_field_writer.add_val(unordered_term_id);
+            }
+
             num_tokens += 1;
         });
 
@@ -183,7 +227,13 @@ impl<Rec: Recorder> SpecializedPostingsWriter<Rec> {
 }
 
 impl<Rec: Recorder> PostingsWriter for SpecializedPostingsWriter<Rec> {
-    fn subscribe(&mut self, doc: DocId, position: u32, term: &Term, ctx: &mut IndexingContext) {
+    fn subscribe(
+        &mut self,
+        doc: DocId,
+        position: u32,
+        term: &Term,
+        ctx: &mut IndexingContext,
+    ) -> UnorderedTermId {
         debug_assert!(term.as_slice().len() >= 4);
         self.total_num_tokens += 1;
         let (term_index, arena) = (&mut ctx.term_index, &mut ctx.arena);
@@ -202,18 +252,18 @@ impl<Rec: Recorder> PostingsWriter for SpecializedPostingsWriter<Rec> {
                 recorder.record_position(position, arena);
                 recorder
             }
-        });
+        }) as UnorderedTermId
     }
 
     fn serialize(
         &self,
-        term_addrs: &[(Term<&[u8]>, Addr)],
+        term_addrs: &[(Term<&[u8]>, Addr, UnorderedTermId)],
         doc_id_map: Option<&DocIdMapping>,
         ctx: &IndexingContext,
         serializer: &mut FieldSerializer,
     ) -> io::Result<()> {
         let mut buffer_lender = BufferLender::default();
-        for (term, addr) in term_addrs {
+        for (term, addr, _) in term_addrs {
             Self::serialize_one_term(term, *addr, doc_id_map, &mut buffer_lender, ctx, serializer)?;
         }
         Ok(())

src/query/mod.rs

@@ -15,7 +15,7 @@ mod more_like_this;
 mod phrase_query;
 mod query;
 mod query_parser;
-// mod range_query;
+mod range_query;
 mod regex_query;
 mod reqopt_scorer;
 mod scorer;
@@ -50,7 +50,7 @@ pub use self::more_like_this::{MoreLikeThisQuery, MoreLikeThisQueryBuilder};
 pub use self::phrase_query::PhraseQuery;
 pub use self::query::{EnableScoring, Query, QueryClone};
 pub use self::query_parser::{QueryParser, QueryParserError};
-// pub use self::range_query::RangeQuery;
+pub use self::range_query::RangeQuery;
 pub use self::regex_query::RegexQuery;
 pub use self::reqopt_scorer::RequiredOptionalScorer;
 pub use self::score_combiner::{

src/query/query_parser/query_parser.rs

@@ -13,19 +13,10 @@ use crate::core::Index;
 use crate::indexer::{
     convert_to_fast_value_and_get_term, set_string_and_get_terms, JsonTermWriter,
 };
-// use crate::query::range_query::is_type_valid_for_fastfield_range_query;
+use crate::query::range_query::is_type_valid_for_fastfield_range_query;
 use crate::query::{
-    AllQuery,
-    BooleanQuery,
-    BoostQuery,
-    EmptyQuery,
-    FuzzyTermQuery,
-    Occur,
-    PhraseQuery,
-    Query,
-    // RangeQuery,
-    TermQuery,
-    TermSetQuery,
+    AllQuery, BooleanQuery, BoostQuery, EmptyQuery, FuzzyTermQuery, Occur, PhraseQuery, Query,
+    RangeQuery, TermQuery, TermSetQuery,
 };
 use crate::schema::{
     Facet, FacetParseError, Field, FieldType, IndexRecordOption, IntoIpv6Addr, JsonObjectOptions,
@@ -343,90 +334,89 @@ impl QueryParser {
         json_path: &str,
         phrase: &str,
     ) -> Result<Term, QueryParserError> {
-        todo!();
-        // let field_entry = self.schema.get_field_entry(field);
-        // let field_type = field_entry.field_type();
-        // let field_supports_ff_range_queries = field_type.is_fast()
-        // && is_type_valid_for_fastfield_range_query(field_type.value_type());
-        //
-        // if !field_type.is_indexed() && !field_supports_ff_range_queries {
-        // return Err(QueryParserError::FieldNotIndexed(
-        // field_entry.name().to_string(),
-        // ));
-        // }
-        // if !json_path.is_empty() && field_type.value_type() != Type::Json {
-        // return Err(QueryParserError::UnsupportedQuery(format!(
-        // "Json path is not supported for field {:?}",
-        // field_entry.name()
-        // )));
-        // }
-        // match *field_type {
-        // FieldType::U64(_) => {
-        // let val: u64 = u64::from_str(phrase)?;
-        // Ok(Term::from_field_u64(field, val))
-        // }
-        // FieldType::I64(_) => {
-        // let val: i64 = i64::from_str(phrase)?;
-        // Ok(Term::from_field_i64(field, val))
-        // }
-        // FieldType::F64(_) => {
-        // let val: f64 = f64::from_str(phrase)?;
-        // Ok(Term::from_field_f64(field, val))
-        // }
-        // FieldType::Bool(_) => {
-        // let val: bool = bool::from_str(phrase)?;
-        // Ok(Term::from_field_bool(field, val))
-        // }
-        // FieldType::Date(_) => {
-        // let dt = OffsetDateTime::parse(phrase, &Rfc3339)?;
-        // Ok(Term::from_field_date(field, DateTime::from_utc(dt)))
-        // }
-        // FieldType::Str(ref str_options) => {
-        // let option = str_options.get_indexing_options().ok_or_else(|| {
-        // This should have been seen earlier really.
-        // QueryParserError::FieldNotIndexed(field_entry.name().to_string())
-        // })?;
-        // let text_analyzer =
-        // self.tokenizer_manager
-        // .get(option.tokenizer())
-        // .ok_or_else(|| QueryParserError::UnknownTokenizer {
-        // field: field_entry.name().to_string(),
-        // tokenizer: option.tokenizer().to_string(),
-        // })?;
-        // let mut terms: Vec<Term> = Vec::new();
-        // let mut token_stream = text_analyzer.token_stream(phrase);
-        // token_stream.process(&mut |token| {
-        // let term = Term::from_field_text(field, &token.text);
-        // terms.push(term);
-        // });
-        // if terms.len() != 1 {
-        // return Err(QueryParserError::UnsupportedQuery(format!(
-        // "Range query boundary cannot have multiple tokens: {phrase:?}."
-        // )));
-        // }
-        // Ok(terms.into_iter().next().unwrap())
-        // }
-        // FieldType::JsonObject(_) => {
-        // Json range are not supported.
-        // Err(QueryParserError::UnsupportedQuery(
-        // "Range query are not supported on json field.".to_string(),
-        // ))
-        // }
-        // FieldType::Facet(_) => match Facet::from_text(phrase) {
-        // Ok(facet) => Ok(Term::from_facet(field, &facet)),
-        // Err(e) => Err(QueryParserError::from(e)),
-        // },
-        // FieldType::Bytes(_) => {
-        // let bytes = BASE64
-        // .decode(phrase)
-        // .map_err(QueryParserError::ExpectedBase64)?;
-        // Ok(Term::from_field_bytes(field, &bytes))
-        // }
-        // FieldType::IpAddr(_) => {
-        // let ip_v6 = IpAddr::from_str(phrase)?.into_ipv6_addr();
-        // Ok(Term::from_field_ip_addr(field, ip_v6))
-        // }
-        // }
+        let field_entry = self.schema.get_field_entry(field);
+        let field_type = field_entry.field_type();
+        let field_supports_ff_range_queries = field_type.is_fast()
+            && is_type_valid_for_fastfield_range_query(field_type.value_type());
+
+        if !field_type.is_indexed() && !field_supports_ff_range_queries {
+            return Err(QueryParserError::FieldNotIndexed(
+                field_entry.name().to_string(),
+            ));
+        }
+        if !json_path.is_empty() && field_type.value_type() != Type::Json {
+            return Err(QueryParserError::UnsupportedQuery(format!(
+                "Json path is not supported for field {:?}",
+                field_entry.name()
+            )));
+        }
+        match *field_type {
+            FieldType::U64(_) => {
+                let val: u64 = u64::from_str(phrase)?;
+                Ok(Term::from_field_u64(field, val))
+            }
+            FieldType::I64(_) => {
+                let val: i64 = i64::from_str(phrase)?;
+                Ok(Term::from_field_i64(field, val))
+            }
+            FieldType::F64(_) => {
+                let val: f64 = f64::from_str(phrase)?;
+                Ok(Term::from_field_f64(field, val))
+            }
+            FieldType::Bool(_) => {
+                let val: bool = bool::from_str(phrase)?;
+                Ok(Term::from_field_bool(field, val))
+            }
+            FieldType::Date(_) => {
+                let dt = OffsetDateTime::parse(phrase, &Rfc3339)?;
+                Ok(Term::from_field_date(field, DateTime::from_utc(dt)))
+            }
+            FieldType::Str(ref str_options) => {
+                let option = str_options.get_indexing_options().ok_or_else(|| {
+                    // This should have been seen earlier really.
+                    QueryParserError::FieldNotIndexed(field_entry.name().to_string())
+                })?;
+                let text_analyzer =
+                    self.tokenizer_manager
+                        .get(option.tokenizer())
+                        .ok_or_else(|| QueryParserError::UnknownTokenizer {
+                            field: field_entry.name().to_string(),
+                            tokenizer: option.tokenizer().to_string(),
+                        })?;
+                let mut terms: Vec<Term> = Vec::new();
+                let mut token_stream = text_analyzer.token_stream(phrase);
+                token_stream.process(&mut |token| {
+                    let term = Term::from_field_text(field, &token.text);
+                    terms.push(term);
+                });
+                if terms.len() != 1 {
+                    return Err(QueryParserError::UnsupportedQuery(format!(
+                        "Range query boundary cannot have multiple tokens: {phrase:?}."
+                    )));
+                }
+                Ok(terms.into_iter().next().unwrap())
+            }
+            FieldType::JsonObject(_) => {
+                // Json range are not supported.
+                Err(QueryParserError::UnsupportedQuery(
+                    "Range query are not supported on json field.".to_string(),
+                ))
+            }
+            FieldType::Facet(_) => match Facet::from_text(phrase) {
+                Ok(facet) => Ok(Term::from_facet(field, &facet)),
+                Err(e) => Err(QueryParserError::from(e)),
+            },
+            FieldType::Bytes(_) => {
+                let bytes = BASE64
+                    .decode(phrase)
+                    .map_err(QueryParserError::ExpectedBase64)?;
+                Ok(Term::from_field_bytes(field, &bytes))
+            }
+            FieldType::IpAddr(_) => {
+                let ip_v6 = IpAddr::from_str(phrase)?.into_ipv6_addr();
+                Ok(Term::from_field_ip_addr(field, ip_v6))
+            }
+        }
     }
 
     fn compute_logical_ast_for_leaf(
@@ -750,12 +740,9 @@ fn convert_literal_to_query(
             value_type,
             lower,
             upper,
-        } => {
-            todo!();
-            // Box::new(RangeQuery::new_term_bounds(
-            //             field, value_type, &lower, &upper,
-            //         ))
-        }
+        } => Box::new(RangeQuery::new_term_bounds(
+            field, value_type, &lower, &upper,
+        )),
         LogicalLiteral::Set { elements, .. } => Box::new(TermSetQuery::new(elements)),
         LogicalLiteral::All => Box::new(AllQuery),
     }

src/query/range_query/fast_field_range_query.rs

@@ -4,7 +4,7 @@ use std::sync::Arc;
 
 use fastfield_codecs::Column;
 
-use crate::fastfield::MakeZero;
+use crate::fastfield::{MakeZero, MultiValuedFastFieldReader};
 use crate::{DocId, DocSet, TERMINATED};
 
 /// Helper to have a cursor over a vec of docids

src/query/range_query/range_query_ip_fastfield.rs

@@ -8,13 +8,10 @@ use std::ops::{Bound, RangeInclusive};
 use common::BinarySerializable;
 use fastfield_codecs::MonotonicallyMappableToU128;
 
+use super::fast_field_range_query::{FastFieldCardinality, RangeDocSet};
 use super::range_query::map_bound;
 use crate::query::{ConstScorer, Explanation, Scorer, Weight};
-<<<<<<< HEAD
 use crate::schema::Cardinality;
-=======
-use crate::schema::Field;
->>>>>>> fd1deefd12 (Disconnected facet / fast field merges / examples)
 use crate::{DocId, DocSet, Score, SegmentReader, TantivyError};
 
 /// `IPFastFieldRangeWeight` uses the ip address fast field to execute range queries.
@@ -43,7 +40,6 @@ impl IPFastFieldRangeWeight {
 
 impl Weight for IPFastFieldRangeWeight {
     fn scorer(&self, reader: &SegmentReader, boost: Score) -> crate::Result<Box<dyn Scorer>> {
-<<<<<<< HEAD
         let field_type = reader
             .schema()
             .get_field_entry(reader.schema().get_field(&self.field)?)
@@ -78,40 +74,6 @@ impl Weight for IPFastFieldRangeWeight {
                 Ok(Box::new(ConstScorer::new(docset, boost)))
             }
         }
-=======
-        todo!();
-        // let field_type = reader.schema().get_field_entry(self.field).field_type();
-        // match field_type.fastfield_cardinality().unwrap() {
-        //     Cardinality::SingleValue => {
-        //         let ip_addr_fast_field = reader.fast_fields().ip_addr(self.field)?;
-        //         let value_range = bound_to_value_range(
-        //             &self.left_bound,
-        //             &self.right_bound,
-        //             ip_addr_fast_field.min_value(),
-        //             ip_addr_fast_field.max_value(),
-        //         );
-        //         let docset = RangeDocSet::new(
-        //             value_range,
-        //             FastFieldCardinality::SingleValue(ip_addr_fast_field),
-        //         );
-        //         Ok(Box::new(ConstScorer::new(docset, boost)))
-        //     }
-        //     Cardinality::MultiValues => {
-        //         let ip_addr_fast_field = reader.fast_fields().ip_addrs(self.field)?;
-        //         let value_range = bound_to_value_range(
-        //             &self.left_bound,
-        //             &self.right_bound,
-        //             ip_addr_fast_field.min_value(),
-        //             ip_addr_fast_field.max_value(),
-        //         );
-        //         let docset = RangeDocSet::new(
-        //             value_range,
-        //             FastFieldCardinality::MultiValue(ip_addr_fast_field),
-        //         );
-        //         Ok(Box::new(ConstScorer::new(docset, boost)))
-        //     }
-        // }
->>>>>>> fd1deefd12 (Disconnected facet / fast field merges / examples)
     }
 
     fn explain(&self, reader: &SegmentReader, doc: DocId) -> crate::Result<Explanation> {
@@ -228,7 +190,7 @@ mod tests {
         let ips_field = schema_builder.add_ip_addr_field(
             "ips",
             IpAddrOptions::default()
-                .set_fast()
+                .set_fast(Cardinality::MultiValues)
                 .set_indexed(),
         );
         let text_field = schema_builder.add_text_field("id", STRING | STORED);

src/query/range_query/range_query_u64_fastfield.rs

@@ -6,14 +6,10 @@ use std::ops::{Bound, RangeInclusive};
 
 use fastfield_codecs::MonotonicallyMappableToU64;
 
-use super::fast_field_range_query::RangeDocSet;
+use super::fast_field_range_query::{FastFieldCardinality, RangeDocSet};
 use super::range_query::map_bound;
 use crate::query::{ConstScorer, Explanation, Scorer, Weight};
-<<<<<<< HEAD
 use crate::schema::Cardinality;
-=======
-use crate::schema::Field;
->>>>>>> fd1deefd12 (Disconnected facet / fast field merges / examples)
 use crate::{DocId, DocSet, Score, SegmentReader, TantivyError};
 
 /// `FastFieldRangeWeight` uses the fast field to execute range queries.
@@ -37,7 +33,6 @@ impl FastFieldRangeWeight {
 
 impl Weight for FastFieldRangeWeight {
     fn scorer(&self, reader: &SegmentReader, boost: Score) -> crate::Result<Box<dyn Scorer>> {
-<<<<<<< HEAD
         let field_type = reader
             .schema()
             .get_field_entry(reader.schema().get_field(&self.field)?)
@@ -68,36 +63,6 @@ impl Weight for FastFieldRangeWeight {
                 Ok(Box::new(ConstScorer::new(docset, boost)))
             }
         }
-=======
-        todo!();
-        // let field_type = reader.schema().get_field_entry(self.field).field_type();
-        // match field_type.fastfield_cardinality().unwrap() {
-        //     Cardinality::SingleValue => {
-        //         let fast_field = reader.fast_fields().u64_lenient(self.field)?;
-        //         let value_range = bound_to_value_range(
-        //             &self.left_bound,
-        //             &self.right_bound,
-        //             fast_field.min_value(),
-        //             fast_field.max_value(),
-        //         );
-        //         let docset =
-        //             RangeDocSet::new(value_range, FastFieldCardinality::SingleValue(fast_field));
-        //         Ok(Box::new(ConstScorer::new(docset, boost)))
-        //     }
-        //     Cardinality::MultiValues => {
-        //         let fast_field = reader.fast_fields().u64s_lenient(self.field)?;
-        //         let value_range = bound_to_value_range(
-        //             &self.left_bound,
-        //             &self.right_bound,
-        //             fast_field.min_value(),
-        //             fast_field.max_value(),
-        //         );
-        //         let docset =
-        //             RangeDocSet::new(value_range, FastFieldCardinality::MultiValue(fast_field));
-        //         Ok(Box::new(ConstScorer::new(docset, boost)))
-        //     }
-        // }
->>>>>>> fd1deefd12 (Disconnected facet / fast field merges / examples)
     }
 
     fn explain(&self, reader: &SegmentReader, doc: DocId) -> crate::Result<Explanation> {
@@ -221,7 +186,7 @@ mod tests {
         let ids_u64_field = schema_builder.add_u64_field(
             "ids",
             NumericOptions::default()
-                .set_fast()
+                .set_fast(Cardinality::MultiValues)
                 .set_indexed(),
         );
 
@@ -229,7 +194,7 @@ mod tests {
         let ids_f64_field = schema_builder.add_f64_field(
             "ids_f64",
             NumericOptions::default()
-                .set_fast()
+                .set_fast(Cardinality::MultiValues)
                 .set_indexed(),
         );
 
@@ -237,7 +202,7 @@ mod tests {
         let ids_i64_field = schema_builder.add_i64_field(
             "ids_i64",
             NumericOptions::default()
-                .set_fast()
+                .set_fast(Cardinality::MultiValues)
                 .set_indexed(),
         );
 

src/schema/date_time_options.rs

@@ -2,16 +2,14 @@ use std::ops::BitOr;
 
 use serde::{Deserialize, Serialize};
 
+use super::Cardinality;
 use crate::schema::flags::{FastFlag, IndexedFlag, SchemaFlagList, StoredFlag};
 
 /// DateTime Precision
-#[derive(
-    Clone, Copy, Debug, Hash, PartialEq, Eq, PartialOrd, Ord, Serialize, Deserialize, Default,
-)]
+#[derive(Clone, Copy, Debug, Hash, PartialEq, Eq, PartialOrd, Ord, Serialize, Deserialize)]
 #[serde(rename_all = "lowercase")]
 pub enum DatePrecision {
     /// Seconds precision
-    #[default]
     Seconds,
     /// Milli-seconds precision.
     Milliseconds,
@@ -19,13 +17,20 @@ pub enum DatePrecision {
     Microseconds,
 }
 
+impl Default for DatePrecision {
+    fn default() -> Self {
+        DatePrecision::Seconds
+    }
+}
+
 /// Defines how DateTime field should be handled by tantivy.
 #[derive(Clone, Debug, PartialEq, Eq, Serialize, Deserialize, Default)]
 pub struct DateOptions {
     indexed: bool,
     // This boolean has no effect if the field is not marked as indexed true.
     fieldnorms: bool,
-    fast: bool,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    fast: Option<Cardinality>,
     stored: bool,
     // Internal storage precision, used to optimize storage
     // compression on fast fields.
@@ -49,9 +54,18 @@ impl DateOptions {
         self.fieldnorms && self.indexed
     }
 
+    /// Returns true iff the value is a fast field and multivalue.
+    pub fn is_multivalue_fast(&self) -> bool {
+        if let Some(cardinality) = self.fast {
+            cardinality == Cardinality::MultiValues
+        } else {
+            false
+        }
+    }
+
     /// Returns true iff the value is a fast field.
     pub fn is_fast(&self) -> bool {
-        self.fast
+        self.fast.is_some()
     }
 
     /// Set the field as stored.
@@ -93,11 +107,19 @@ impl DateOptions {
     /// If more than one value is associated with a fast field, only the last one is
     /// kept.
     #[must_use]
-    pub fn set_fast(mut self) -> DateOptions {
-        self.fast = true;
+    pub fn set_fast(mut self, cardinality: Cardinality) -> DateOptions {
+        self.fast = Some(cardinality);
         self
     }
 
+    /// Returns the cardinality of the fastfield.
+    ///
+    /// If the field has not been declared as a fastfield, then
+    /// the method returns `None`.
+    pub fn get_fastfield_cardinality(&self) -> Option<Cardinality> {
+        self.fast
+    }
+
     /// Sets the precision for this DateTime field.
     ///
     /// Internal storage precision, used to optimize storage
@@ -125,7 +147,10 @@ impl From<()> for DateOptions {
 impl From<FastFlag> for DateOptions {
     fn from(_: FastFlag) -> Self {
         DateOptions {
-            fast: true,
+            indexed: false,
+            fieldnorms: false,
+            stored: false,
+            fast: Some(Cardinality::SingleValue),
             ..Default::default()
         }
     }
@@ -134,7 +159,10 @@ impl From<FastFlag> for DateOptions {
 impl From<StoredFlag> for DateOptions {
     fn from(_: StoredFlag) -> Self {
         DateOptions {
+            indexed: false,
+            fieldnorms: false,
             stored: true,
+            fast: None,
             ..Default::default()
         }
     }
@@ -145,6 +173,8 @@ impl From<IndexedFlag> for DateOptions {
         DateOptions {
             indexed: true,
             fieldnorms: true,
+            stored: false,
+            fast: None,
             ..Default::default()
         }
     }
@@ -159,7 +189,7 @@ impl<T: Into<DateOptions>> BitOr<T> for DateOptions {
             indexed: self.indexed | other.indexed,
             fieldnorms: self.fieldnorms | other.fieldnorms,
             stored: self.stored | other.stored,
-            fast: self.fast | other.fast,
+            fast: self.fast.or(other.fast),
             precision: self.precision,
         }
     }

src/schema/field_type.rs

@@ -8,7 +8,7 @@ use serde_json::Value as JsonValue;
 use thiserror::Error;
 
 use super::ip_options::IpAddrOptions;
-use super::IntoIpv6Addr;
+use super::{Cardinality, IntoIpv6Addr};
 use crate::schema::bytes_options::BytesOptions;
 use crate::schema::facet_options::FacetOptions;
 use crate::schema::{
@@ -241,6 +241,26 @@ impl FieldType {
         }
     }
 
+    /// returns true if the field is fast.
+    pub fn fastfield_cardinality(&self) -> Option<Cardinality> {
+        match *self {
+            FieldType::Bytes(ref bytes_options) => {
+                bytes_options.is_fast().then_some(Cardinality::SingleValue)
+            }
+            FieldType::Str(ref text_options) => {
+                text_options.is_fast().then_some(Cardinality::MultiValues)
+            }
+            FieldType::U64(ref int_options)
+            | FieldType::I64(ref int_options)
+            | FieldType::F64(ref int_options)
+            | FieldType::Bool(ref int_options) => int_options.get_fastfield_cardinality(),
+            FieldType::Date(ref date_options) => date_options.get_fastfield_cardinality(),
+            FieldType::Facet(_) => Some(Cardinality::MultiValues),
+            FieldType::JsonObject(_) => None,
+            FieldType::IpAddr(ref ip_addr_options) => ip_addr_options.get_fastfield_cardinality(),
+        }
+    }
+
     /// returns true if the field is normed (see [fieldnorms](crate::fieldnorm)).
     pub fn has_fieldnorms(&self) -> bool {
         match *self {

src/schema/ip_options.rs

@@ -4,6 +4,7 @@ use std::ops::BitOr;
 use serde::{Deserialize, Serialize};
 
 use super::flags::{FastFlag, IndexedFlag, SchemaFlagList, StoredFlag};
+use super::Cardinality;
 
 /// Trait to convert into an Ipv6Addr.
 pub trait IntoIpv6Addr {
@@ -23,7 +24,8 @@ impl IntoIpv6Addr for IpAddr {
 /// Define how an ip field should be handled by tantivy.
 #[derive(Clone, Debug, PartialEq, Eq, Serialize, Deserialize, Default)]
 pub struct IpAddrOptions {
-    fast: bool,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    fast: Option<Cardinality>,
     stored: bool,
     indexed: bool,
     fieldnorms: bool,
@@ -32,7 +34,7 @@ pub struct IpAddrOptions {
 impl IpAddrOptions {
     /// Returns true iff the value is a fast field.
     pub fn is_fast(&self) -> bool {
-        self.fast
+        self.fast.is_some()
     }
 
     /// Returns `true` if the ip address should be stored in the doc store.
@@ -50,6 +52,14 @@ impl IpAddrOptions {
         self.fieldnorms
     }
 
+    /// Returns the cardinality of the fastfield.
+    ///
+    /// If the field has not been declared as a fastfield, then
+    /// the method returns None.
+    pub fn get_fastfield_cardinality(&self) -> Option<Cardinality> {
+        self.fast
+    }
+
     /// Set the field as normed.
     ///
     /// Setting an integer as normed will generate
@@ -87,8 +97,8 @@ impl IpAddrOptions {
     /// If more than one value is associated with a fast field, only the last one is
     /// kept.
     #[must_use]
-    pub fn set_fast(mut self) -> Self {
-        self.fast = true;
+    pub fn set_fast(mut self, cardinality: Cardinality) -> Self {
+        self.fast = Some(cardinality);
         self
     }
 }
@@ -105,7 +115,7 @@ impl From<FastFlag> for IpAddrOptions {
             fieldnorms: false,
             indexed: false,
             stored: false,
-            fast: true,
+            fast: Some(Cardinality::SingleValue),
         }
     }
 }
@@ -116,7 +126,7 @@ impl From<StoredFlag> for IpAddrOptions {
             fieldnorms: false,
             indexed: false,
             stored: true,
-            fast: false,
+            fast: None,
         }
     }
 }
@@ -127,7 +137,7 @@ impl From<IndexedFlag> for IpAddrOptions {
             fieldnorms: true,
             indexed: true,
             stored: false,
-            fast: false,
+            fast: None,
         }
     }
 }
@@ -141,7 +151,7 @@ impl<T: Into<IpAddrOptions>> BitOr<T> for IpAddrOptions {
             fieldnorms: self.fieldnorms | other.fieldnorms,
             indexed: self.indexed | other.indexed,
             stored: self.stored | other.stored,
-            fast: self.fast | other.fast,
+            fast: self.fast.or(other.fast),
         }
     }
 }

src/schema/mod.rs

@@ -141,9 +141,9 @@ pub use self::index_record_option::IndexRecordOption;
 pub use self::ip_options::{IntoIpv6Addr, IpAddrOptions};
 pub use self::json_object_options::JsonObjectOptions;
 pub use self::named_field_document::NamedFieldDocument;
-#[allow(deprecated)]
-pub use self::numeric_options::IntOptions;
 pub use self::numeric_options::NumericOptions;
+#[allow(deprecated)]
+pub use self::numeric_options::{Cardinality, IntOptions};
 pub use self::schema::{DocParsingError, Schema, SchemaBuilder};
 pub use self::term::Term;
 pub use self::text_options::{TextFieldIndexing, TextOptions, STRING, TEXT};

src/schema/numeric_options.rs

@@ -4,6 +4,18 @@ use serde::{Deserialize, Serialize};
 
 use crate::schema::flags::{FastFlag, IndexedFlag, SchemaFlagList, StoredFlag};
 
+/// Express whether a field is single-value or multi-valued.
+#[derive(Clone, Copy, PartialEq, Eq, Debug, Serialize, Deserialize)]
+pub enum Cardinality {
+    /// The document must have exactly one value associated with the document.
+    #[serde(rename = "single")]
+    SingleValue,
+    /// The document can have any number of values associated with the document.
+    /// This is more memory and CPU expensive than the `SingleValue` solution.
+    #[serde(rename = "multi")]
+    MultiValues,
+}
+
 #[deprecated(since = "0.17.0", note = "Use NumericOptions instead.")]
 /// Deprecated use [`NumericOptions`] instead.
 pub type IntOptions = NumericOptions;
@@ -15,7 +27,8 @@ pub struct NumericOptions {
     indexed: bool,
     // This boolean has no effect if the field is not marked as indexed too.
     fieldnorms: bool, // This attribute only has an effect if indexed is true.
-    fast: bool,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    fast: Option<Cardinality>,
     stored: bool,
 }
 
@@ -29,7 +42,8 @@ struct NumericOptionsDeser {
     indexed: bool,
     #[serde(default)]
     fieldnorms: Option<bool>, // This attribute only has an effect if indexed is true.
-    fast: bool,
+    #[serde(default)]
+    fast: Option<Cardinality>,
     stored: bool,
 }
 
@@ -60,9 +74,18 @@ impl NumericOptions {
         self.fieldnorms && self.indexed
     }
 
+    /// Returns true iff the value is a fast field and multivalue.
+    pub fn is_multivalue_fast(&self) -> bool {
+        if let Some(cardinality) = self.fast {
+            cardinality == Cardinality::MultiValues
+        } else {
+            false
+        }
+    }
+
     /// Returns true iff the value is a fast field.
     pub fn is_fast(&self) -> bool {
-        self.fast
+        self.fast.is_some()
     }
 
     /// Set the field as stored.
@@ -104,10 +127,18 @@ impl NumericOptions {
     /// If more than one value is associated with a fast field, only the last one is
     /// kept.
     #[must_use]
-    pub fn set_fast(mut self) -> NumericOptions {
-        self.fast = true;
+    pub fn set_fast(mut self, cardinality: Cardinality) -> NumericOptions {
+        self.fast = Some(cardinality);
         self
     }
+
+    /// Returns the cardinality of the fastfield.
+    ///
+    /// If the field has not been declared as a fastfield, then
+    /// the method returns `None`.
+    pub fn get_fastfield_cardinality(&self) -> Option<Cardinality> {
+        self.fast
+    }
 }
 
 impl From<()> for NumericOptions {
@@ -122,7 +153,7 @@ impl From<FastFlag> for NumericOptions {
             indexed: false,
             fieldnorms: false,
             stored: false,
-            fast: true,
+            fast: Some(Cardinality::SingleValue),
         }
     }
 }
@@ -133,7 +164,7 @@ impl From<StoredFlag> for NumericOptions {
             indexed: false,
             fieldnorms: false,
             stored: true,
-            fast: false,
+            fast: None,
         }
     }
 }
@@ -144,7 +175,7 @@ impl From<IndexedFlag> for NumericOptions {
             indexed: true,
             fieldnorms: true,
             stored: false,
-            fast: false,
+            fast: None,
         }
     }
 }
@@ -158,7 +189,7 @@ impl<T: Into<NumericOptions>> BitOr<T> for NumericOptions {
             indexed: self.indexed | other.indexed,
             fieldnorms: self.fieldnorms | other.fieldnorms,
             stored: self.stored | other.stored,
-            fast: self.fast | other.fast,
+            fast: self.fast.or(other.fast),
         }
     }
 }
@@ -190,7 +221,7 @@ mod tests {
             &NumericOptions {
                 indexed: true,
                 fieldnorms: true,
-                fast: false,
+                fast: None,
                 stored: false
             }
         );
@@ -208,7 +239,7 @@ mod tests {
             &NumericOptions {
                 indexed: false,
                 fieldnorms: false,
-                fast: false,
+                fast: None,
                 stored: false
             }
         );
@@ -227,7 +258,7 @@ mod tests {
             &NumericOptions {
                 indexed: true,
                 fieldnorms: false,
-                fast: false,
+                fast: None,
                 stored: false
             }
         );
@@ -247,7 +278,7 @@ mod tests {
             &NumericOptions {
                 indexed: false,
                 fieldnorms: true,
-                fast: false,
+                fast: None,
                 stored: false
             }
         );

src/schema/schema.rs

@@ -484,6 +484,7 @@ mod tests {
     use serde_json;
 
     use crate::schema::field_type::ValueParsingError;
+    use crate::schema::numeric_options::Cardinality::SingleValue;
     use crate::schema::schema::DocParsingError::InvalidJson;
     use crate::schema::*;
 
@@ -505,13 +506,19 @@ mod tests {
     #[test]
     pub fn test_schema_serialization() {
         let mut schema_builder = Schema::builder();
-        let count_options = NumericOptions::default().set_stored().set_fast();
-        let popularity_options = NumericOptions::default().set_stored().set_fast();
+        let count_options = NumericOptions::default()
+            .set_stored()
+            .set_fast(Cardinality::SingleValue);
+        let popularity_options = NumericOptions::default()
+            .set_stored()
+            .set_fast(Cardinality::SingleValue);
         let score_options = NumericOptions::default()
             .set_indexed()
             .set_fieldnorm()
-            .set_fast();
-        let is_read_options = NumericOptions::default().set_stored().set_fast();
+            .set_fast(Cardinality::SingleValue);
+        let is_read_options = NumericOptions::default()
+            .set_stored()
+            .set_fast(Cardinality::SingleValue);
         schema_builder.add_text_field("title", TEXT);
         schema_builder.add_text_field(
             "author",
@@ -636,8 +643,12 @@ mod tests {
     #[test]
     pub fn test_document_to_json() {
         let mut schema_builder = Schema::builder();
-        let count_options = NumericOptions::default().set_stored().set_fast();
-        let is_read_options = NumericOptions::default().set_stored().set_fast();
+        let count_options = NumericOptions::default()
+            .set_stored()
+            .set_fast(Cardinality::SingleValue);
+        let is_read_options = NumericOptions::default()
+            .set_stored()
+            .set_fast(Cardinality::SingleValue);
         schema_builder.add_text_field("title", TEXT);
         schema_builder.add_text_field("author", STRING);
         schema_builder.add_u64_field("count", count_options);
@@ -737,9 +748,15 @@ mod tests {
     #[test]
     pub fn test_parse_document() {
         let mut schema_builder = Schema::builder();
-        let count_options = NumericOptions::default().set_stored().set_fast();
-        let popularity_options = NumericOptions::default().set_stored().set_fast();
-        let score_options = NumericOptions::default().set_indexed().set_fast();
+        let count_options = NumericOptions::default()
+            .set_stored()
+            .set_fast(Cardinality::SingleValue);
+        let popularity_options = NumericOptions::default()
+            .set_stored()
+            .set_fast(Cardinality::SingleValue);
+        let score_options = NumericOptions::default()
+            .set_indexed()
+            .set_fast(Cardinality::SingleValue);
         let title_field = schema_builder.add_text_field("title", TEXT);
         let author_field = schema_builder.add_text_field("author", STRING);
         let count_field = schema_builder.add_u64_field("count", count_options);
@@ -890,7 +907,7 @@ mod tests {
             .set_stored()
             .set_indexed()
             .set_fieldnorm()
-            .set_fast();
+            .set_fast(SingleValue);
         schema_builder.add_text_field("_id", id_options);
         schema_builder.add_date_field("_timestamp", timestamp_options);