make convert_to_fast_value_and_append_to_json_term pub

* remove JsonTermWriter

remove JsonTermWriter
remove path truncation logic, add assertion

* fix json_path_writer add sep logic

CHANGELOG.md

@@ -1,4 +1,4 @@
-Tantivy 0.22 (unreleased)
+Tantivy 0.22
 ================================
 
 Tantivy 0.22 will be able to read indices created with Tantivy 0.21.

common/src/bitset.rs

@@ -1,5 +1,5 @@
 use std::io::Write;
-use std::{fmt, io, u64};
+use std::{fmt, io};
 
 use ownedbytes::OwnedBytes;
 

common/src/json_path_writer.rs

@@ -5,6 +5,12 @@ pub const JSON_PATH_SEGMENT_SEP: u8 = 1u8;
 pub const JSON_PATH_SEGMENT_SEP_STR: &str =
     unsafe { std::str::from_utf8_unchecked(&[JSON_PATH_SEGMENT_SEP]) };
 
+/// Separates the json path and the value in
+/// a JSON term binary representation.
+pub const JSON_END_OF_PATH: u8 = 0u8;
+pub const JSON_END_OF_PATH_STR: &str =
+    unsafe { std::str::from_utf8_unchecked(&[JSON_END_OF_PATH]) };
+
 /// Create a new JsonPathWriter, that creates flattened json paths for tantivy.
 #[derive(Clone, Debug, Default)]
 pub struct JsonPathWriter {
@@ -14,6 +20,14 @@ pub struct JsonPathWriter {
 }
 
 impl JsonPathWriter {
+    pub fn with_expand_dots(expand_dots: bool) -> Self {
+        JsonPathWriter {
+            path: String::new(),
+            indices: Vec::new(),
+            expand_dots,
+        }
+    }
+
     pub fn new() -> Self {
         JsonPathWriter {
             path: String::new(),
@@ -39,8 +53,8 @@ impl JsonPathWriter {
     pub fn push(&mut self, segment: &str) {
         let len_path = self.path.len();
         self.indices.push(len_path);
-        if !self.path.is_empty() {
-            self.path.push_str(JSON_PATH_SEGMENT_SEP_STR);
+        if self.indices.len() > 1 {
+            self.path.push(JSON_PATH_SEGMENT_SEP as char);
         }
         self.path.push_str(segment);
         if self.expand_dots {
@@ -55,6 +69,12 @@ impl JsonPathWriter {
         }
     }
 
+    /// Set the end of JSON path marker.
+    #[inline]
+    pub fn set_end(&mut self) {
+        self.path.push_str(JSON_END_OF_PATH_STR);
+    }
+
     /// Remove the last segment. Does nothing if the path is empty.
     #[inline]
     pub fn pop(&mut self) {
@@ -91,6 +111,7 @@ mod tests {
     #[test]
     fn json_path_writer_test() {
         let mut writer = JsonPathWriter::new();
+        writer.set_expand_dots(false);
 
         writer.push("root");
         assert_eq!(writer.as_str(), "root");
@@ -109,4 +130,15 @@ mod tests {
         writer.push("k8s.node.id");
         assert_eq!(writer.as_str(), "root\u{1}k8s\u{1}node\u{1}id");
     }
+
+    #[test]
+    fn test_json_path_expand_dots_enabled_pop_segment() {
+        let mut json_writer = JsonPathWriter::with_expand_dots(true);
+        json_writer.push("hello");
+        assert_eq!(json_writer.as_str(), "hello");
+        json_writer.push("color.hue");
+        assert_eq!(json_writer.as_str(), "hello\x01color\x01hue");
+        json_writer.pop();
+        assert_eq!(json_writer.as_str(), "hello");
+    }
 }

common/src/lib.rs

@@ -9,7 +9,7 @@ mod byte_count;
 mod datetime;
 pub mod file_slice;
 mod group_by;
-mod json_path_writer;
+pub mod json_path_writer;
 mod serialize;
 mod vint;
 mod writer;

query-grammar/src/query_grammar.rs

@@ -218,27 +218,14 @@ fn term_or_phrase_infallible(inp: &str) -> JResult<&str, Option<UserInputLeaf>>
 }
 
 fn term_group(inp: &str) -> IResult<&str, UserInputAst> {
-    let occur_symbol = alt((
-        value(Occur::MustNot, char('-')),
-        value(Occur::Must, char('+')),
-    ));
-
     map(
         tuple((
             terminated(field_name, multispace0),
-            delimited(
-                tuple((char('('), multispace0)),
-                separated_list0(multispace1, tuple((opt(occur_symbol), term_or_phrase))),
-                char(')'),
-            ),
+            delimited(tuple((char('('), multispace0)), ast, char(')')),
         )),
-        |(field_name, terms)| {
-            UserInputAst::Clause(
-                terms
-                    .into_iter()
-                    .map(|(occur, leaf)| (occur, leaf.set_field(Some(field_name.clone())).into()))
-                    .collect(),
-            )
+        |(field_name, mut ast)| {
+            ast.set_default_field(field_name);
+            ast
         },
     )(inp)
 }
@@ -258,46 +245,18 @@ fn term_group_precond(inp: &str) -> IResult<&str, (), ()> {
 }
 
 fn term_group_infallible(inp: &str) -> JResult<&str, UserInputAst> {
-    let (mut inp, (field_name, _, _, _)) =
+    let (inp, (field_name, _, _, _)) =
         tuple((field_name, multispace0, char('('), multispace0))(inp).expect("precondition failed");
 
-    let mut terms = Vec::new();
-    let mut errs = Vec::new();
-
-    let mut first_round = true;
-    loop {
-        let mut space_error = if first_round {
-            first_round = false;
-            Vec::new()
-        } else {
-            let (rest, (_, err)) = space1_infallible(inp)?;
-            inp = rest;
-            err
-        };
-        if inp.is_empty() {
-            errs.push(LenientErrorInternal {
-                pos: inp.len(),
-                message: "missing )".to_string(),
-            });
-            break Ok((inp, (UserInputAst::Clause(terms), errs)));
-        }
-        if let Some(inp) = inp.strip_prefix(')') {
-            break Ok((inp, (UserInputAst::Clause(terms), errs)));
-        }
-        // only append missing space error if we did not reach the end of group
-        errs.append(&mut space_error);
-
-        // here we do the assumption term_or_phrase_infallible always consume something if the
-        // first byte is not `)` or ' '. If it did not, we would end up looping.
-
-        let (rest, ((occur, leaf), mut err)) =
-            tuple_infallible((occur_symbol, term_or_phrase_infallible))(inp)?;
-        errs.append(&mut err);
-        if let Some(leaf) = leaf {
-            terms.push((occur, leaf.set_field(Some(field_name.clone())).into()));
-        }
-        inp = rest;
-    }
+    let res = delimited_infallible(
+        nothing,
+        map(ast_infallible, |(mut ast, errors)| {
+            ast.set_default_field(field_name.to_string());
+            (ast, errors)
+        }),
+        opt_i_err(char(')'), "expected ')'"),
+    )(inp);
+    res
 }
 
 fn exists(inp: &str) -> IResult<&str, UserInputLeaf> {
@@ -1468,8 +1427,18 @@ mod test {
 
     #[test]
     fn test_parse_query_term_group() {
-        test_parse_query_to_ast_helper(r#"field:(abc)"#, r#"(*"field":abc)"#);
+        test_parse_query_to_ast_helper(r#"field:(abc)"#, r#""field":abc"#);
         test_parse_query_to_ast_helper(r#"field:(+a -"b c")"#, r#"(+"field":a -"field":"b c")"#);
+        test_parse_query_to_ast_helper(r#"field:(a AND "b c")"#, r#"(+"field":a +"field":"b c")"#);
+        test_parse_query_to_ast_helper(r#"field:(a OR "b c")"#, r#"(?"field":a ?"field":"b c")"#);
+        test_parse_query_to_ast_helper(
+            r#"field:(a OR (b AND c))"#,
+            r#"(?"field":a ?(+"field":b +"field":c))"#,
+        );
+        test_parse_query_to_ast_helper(
+            r#"field:(a [b TO c])"#,
+            r#"(*"field":a *"field":["b" TO "c"])"#,
+        );
 
         test_is_parse_err(r#"field:(+a -"b c""#, r#"(+"field":a -"field":"b c")"#);
     }

query-grammar/src/user_input_ast.rs

@@ -44,6 +44,26 @@ impl UserInputLeaf {
             },
         }
     }
+
+    pub(crate) fn set_default_field(&mut self, default_field: String) {
+        match self {
+            UserInputLeaf::Literal(ref mut literal) if literal.field_name.is_none() => {
+                literal.field_name = Some(default_field)
+            }
+            UserInputLeaf::All => {
+                *self = UserInputLeaf::Exists {
+                    field: default_field,
+                }
+            }
+            UserInputLeaf::Range { ref mut field, .. } if field.is_none() => {
+                *field = Some(default_field)
+            }
+            UserInputLeaf::Set { ref mut field, .. } if field.is_none() => {
+                *field = Some(default_field)
+            }
+            _ => (), // field was already set, do nothing
+        }
+    }
 }
 
 impl Debug for UserInputLeaf {
@@ -205,6 +225,16 @@ impl UserInputAst {
     pub fn or(asts: Vec<UserInputAst>) -> UserInputAst {
         UserInputAst::compose(Occur::Should, asts)
     }
+
+    pub(crate) fn set_default_field(&mut self, field: String) {
+        match self {
+            UserInputAst::Clause(clauses) => clauses
+                .iter_mut()
+                .for_each(|(_, ast)| ast.set_default_field(field.clone())),
+            UserInputAst::Leaf(leaf) => leaf.set_default_field(field),
+            UserInputAst::Boost(ref mut ast, _) => ast.set_default_field(field),
+        }
+    }
 }
 
 impl From<UserInputLiteral> for UserInputLeaf {

src/aggregation/bucket/mod.rs

@@ -28,6 +28,7 @@ mod term_agg;
 mod term_missing_agg;
 
 use std::collections::HashMap;
+use std::fmt;
 
 pub use histogram::*;
 pub use range::*;
@@ -72,12 +73,12 @@ impl From<&str> for OrderTarget {
     }
 }
 
-impl ToString for OrderTarget {
-    fn to_string(&self) -> String {
+impl fmt::Display for OrderTarget {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
         match self {
-            OrderTarget::Key => "_key".to_string(),
-            OrderTarget::Count => "_count".to_string(),
-            OrderTarget::SubAggregation(agg) => agg.to_string(),
+            OrderTarget::Key => f.write_str("_key"),
+            OrderTarget::Count => f.write_str("_count"),
+            OrderTarget::SubAggregation(agg) => agg.fmt(f),
         }
     }
 }

src/aggregation/metric/top_hits.rs

@@ -2,6 +2,7 @@ use std::collections::HashMap;
 use std::net::Ipv6Addr;
 
 use columnar::{ColumnarReader, DynamicColumn};
+use common::json_path_writer::JSON_PATH_SEGMENT_SEP_STR;
 use common::DateTime;
 use regex::Regex;
 use serde::ser::SerializeMap;
@@ -15,7 +16,6 @@ use crate::aggregation::intermediate_agg_result::{
 use crate::aggregation::segment_agg_result::SegmentAggregationCollector;
 use crate::aggregation::AggregationError;
 use crate::collector::TopNComputer;
-use crate::schema::term::JSON_PATH_SEGMENT_SEP_STR;
 use crate::schema::OwnedValue;
 use crate::{DocAddress, DocId, SegmentOrdinal};
 

src/aggregation/mod.rs

@@ -159,10 +159,6 @@ use itertools::Itertools;
 use serde::de::{self, Visitor};
 use serde::{Deserialize, Deserializer, Serialize};
 
-pub(crate) fn invalid_agg_request(message: String) -> crate::TantivyError {
-    crate::TantivyError::AggregationError(AggregationError::InvalidRequest(message))
-}
-
 fn parse_str_into_f64<E: de::Error>(value: &str) -> Result<f64, E> {
     let parsed = value.parse::<f64>().map_err(|_err| {
         de::Error::custom(format!("Failed to parse f64 from string: {:?}", value))

src/collector/facet_collector.rs

@@ -1,7 +1,7 @@
 use std::cmp::Ordering;
 use std::collections::{btree_map, BTreeMap, BTreeSet, BinaryHeap};
+use std::io;
 use std::ops::Bound;
-use std::{io, u64, usize};
 
 use crate::collector::{Collector, SegmentCollector};
 use crate::fastfield::FacetReader;

src/core/json_utils.rs

@@ -1,12 +1,11 @@
-use columnar::MonotonicallyMappableToU64;
+use common::json_path_writer::JSON_PATH_SEGMENT_SEP;
 use common::{replace_in_place, JsonPathWriter};
 use rustc_hash::FxHashMap;
 
-use crate::fastfield::FastValue;
 use crate::postings::{IndexingContext, IndexingPosition, PostingsWriter};
 use crate::schema::document::{ReferenceValue, ReferenceValueLeaf, Value};
-use crate::schema::term::JSON_PATH_SEGMENT_SEP;
-use crate::schema::{Field, Type, DATE_TIME_PRECISION_INDEXED};
+use crate::schema::indexing_term::IndexingTerm;
+use crate::schema::{Field, Type};
 use crate::time::format_description::well_known::Rfc3339;
 use crate::time::{OffsetDateTime, UtcOffset};
 use crate::tokenizer::TextAnalyzer;
@@ -76,7 +75,7 @@ pub(crate) fn index_json_values<'a, V: Value<'a>>(
     json_visitors: impl Iterator<Item = crate::Result<V::ObjectIter>>,
     text_analyzer: &mut TextAnalyzer,
     expand_dots_enabled: bool,
-    term_buffer: &mut Term,
+    term_buffer: &mut IndexingTerm,
     postings_writer: &mut dyn PostingsWriter,
     json_path_writer: &mut JsonPathWriter,
     ctx: &mut IndexingContext,
@@ -105,7 +104,7 @@ fn index_json_object<'a, V: Value<'a>>(
     doc: DocId,
     json_visitor: V::ObjectIter,
     text_analyzer: &mut TextAnalyzer,
-    term_buffer: &mut Term,
+    term_buffer: &mut IndexingTerm,
     json_path_writer: &mut JsonPathWriter,
     postings_writer: &mut dyn PostingsWriter,
     ctx: &mut IndexingContext,
@@ -132,19 +131,16 @@ fn index_json_value<'a, V: Value<'a>>(
     doc: DocId,
     json_value: V,
     text_analyzer: &mut TextAnalyzer,
-    term_buffer: &mut Term,
+    term_buffer: &mut IndexingTerm,
     json_path_writer: &mut JsonPathWriter,
     postings_writer: &mut dyn PostingsWriter,
     ctx: &mut IndexingContext,
     positions_per_path: &mut IndexingPositionsPerPath,
 ) {
-    let set_path_id = |term_buffer: &mut Term, unordered_id: u32| {
+    let set_path_id = |term_buffer: &mut IndexingTerm, unordered_id: u32| {
         term_buffer.truncate_value_bytes(0);
         term_buffer.append_bytes(&unordered_id.to_be_bytes());
     };
-    let set_type = |term_buffer: &mut Term, typ: Type| {
-        term_buffer.append_bytes(&[typ.to_code()]);
-    };
 
     match json_value.as_value() {
         ReferenceValue::Leaf(leaf) => match leaf {
@@ -157,7 +153,7 @@ fn index_json_value<'a, V: Value<'a>>(
 
                 // TODO: make sure the chain position works out.
                 set_path_id(term_buffer, unordered_id);
-                set_type(term_buffer, Type::Str);
+                term_buffer.append_bytes(&[Type::Str.to_code()]);
                 let indexing_position = positions_per_path.get_position_from_id(unordered_id);
                 postings_writer.index_text(
                     doc,
@@ -213,18 +209,16 @@ fn index_json_value<'a, V: Value<'a>>(
                 postings_writer.subscribe(doc, 0u32, term_buffer, ctx);
             }
             ReferenceValueLeaf::PreTokStr(_) => {
-                unimplemented!(
-                    "Pre-tokenized string support in dynamic fields is not yet implemented"
-                )
+                unimplemented!("Pre-tokenized string support in JSON fields is not yet implemented")
             }
             ReferenceValueLeaf::Bytes(_) => {
-                unimplemented!("Bytes support in dynamic fields is not yet implemented")
+                unimplemented!("Bytes support in JSON fields is not yet implemented")
             }
             ReferenceValueLeaf::Facet(_) => {
-                unimplemented!("Facet support in dynamic fields is not yet implemented")
+                unimplemented!("Facet support in JSON fields is not yet implemented")
             }
             ReferenceValueLeaf::IpAddr(_) => {
-                unimplemented!("IP address support in dynamic fields is not yet implemented")
+                unimplemented!("IP address support in JSON fields is not yet implemented")
             }
         },
         ReferenceValue::Array(elements) => {
@@ -256,71 +250,43 @@ fn index_json_value<'a, V: Value<'a>>(
     }
 }
 
-// Tries to infer a JSON type from a string.
-pub fn convert_to_fast_value_and_get_term(
-    json_term_writer: &mut JsonTermWriter,
-    phrase: &str,
-) -> Option<Term> {
+/// Tries to infer a JSON type from a string and append it to the term.
+///
+/// The term must be json + JSON path.
+pub fn convert_to_fast_value_and_append_to_json_term(mut term: Term, phrase: &str) -> Option<Term> {
+    assert_eq!(
+        term.value()
+            .as_json()
+            .expect("expecting a Term with a json type and json path")
+            .1
+            .as_serialized()
+            .len(),
+        0,
+        "JSON value bytes should be empty"
+    );
     if let Ok(dt) = OffsetDateTime::parse(phrase, &Rfc3339) {
         let dt_utc = dt.to_offset(UtcOffset::UTC);
-        return Some(set_fastvalue_and_get_term(
-            json_term_writer,
-            DateTime::from_utc(dt_utc),
-        ));
+        term.append_type_and_fast_value(DateTime::from_utc(dt_utc));
+        return Some(term);
     }
     if let Ok(i64_val) = str::parse::<i64>(phrase) {
-        return Some(set_fastvalue_and_get_term(json_term_writer, i64_val));
+        term.append_type_and_fast_value(i64_val);
+        return Some(term);
     }
     if let Ok(u64_val) = str::parse::<u64>(phrase) {
-        return Some(set_fastvalue_and_get_term(json_term_writer, u64_val));
+        term.append_type_and_fast_value(u64_val);
+        return Some(term);
     }
     if let Ok(f64_val) = str::parse::<f64>(phrase) {
-        return Some(set_fastvalue_and_get_term(json_term_writer, f64_val));
+        term.append_type_and_fast_value(f64_val);
+        return Some(term);
     }
     if let Ok(bool_val) = str::parse::<bool>(phrase) {
-        return Some(set_fastvalue_and_get_term(json_term_writer, bool_val));
+        term.append_type_and_fast_value(bool_val);
+        return Some(term);
     }
     None
 }
-// helper function to generate a Term from a json fastvalue
-pub(crate) fn set_fastvalue_and_get_term<T: FastValue>(
-    json_term_writer: &mut JsonTermWriter,
-    value: T,
-) -> Term {
-    json_term_writer.set_fast_value(value);
-    json_term_writer.term().clone()
-}
-
-// helper function to generate a list of terms with their positions from a textual json value
-pub(crate) fn set_string_and_get_terms(
-    json_term_writer: &mut JsonTermWriter,
-    value: &str,
-    text_analyzer: &mut TextAnalyzer,
-) -> Vec<(usize, Term)> {
-    let mut positions_and_terms = Vec::<(usize, Term)>::new();
-    json_term_writer.close_path_and_set_type(Type::Str);
-    let term_num_bytes = json_term_writer.term_buffer.len_bytes();
-    let mut token_stream = text_analyzer.token_stream(value);
-    token_stream.process(&mut |token| {
-        json_term_writer
-            .term_buffer
-            .truncate_value_bytes(term_num_bytes);
-        json_term_writer
-            .term_buffer
-            .append_bytes(token.text.as_bytes());
-        positions_and_terms.push((token.position, json_term_writer.term().clone()));
-    });
-    positions_and_terms
-}
-
-/// Writes a value of a JSON field to a `Term`.
-/// The Term format is as follows:
-/// `[JSON_TYPE][JSON_PATH][JSON_END_OF_PATH][VALUE_BYTES]`
-pub struct JsonTermWriter<'a> {
-    term_buffer: &'a mut Term,
-    path_stack: Vec<usize>,
-    expand_dots_enabled: bool,
-}
 
 /// Splits a json path supplied to the query parser in such a way that
 /// `.` can be escaped.
@@ -377,275 +343,106 @@ pub(crate) fn encode_column_name(
     path.into()
 }
 
-impl<'a> JsonTermWriter<'a> {
-    pub fn from_field_and_json_path(
-        field: Field,
-        json_path: &str,
-        expand_dots_enabled: bool,
-        term_buffer: &'a mut Term,
-    ) -> Self {
-        term_buffer.set_field_and_type(field, Type::Json);
-        let mut json_term_writer = Self::wrap(term_buffer, expand_dots_enabled);
-        for segment in split_json_path(json_path) {
-            json_term_writer.push_path_segment(&segment);
-        }
-        json_term_writer
+pub fn term_from_json_paths<'a>(
+    json_field: Field,
+    paths: impl Iterator<Item = &'a str>,
+    expand_dots_enabled: bool,
+) -> Term {
+    let mut json_path = JsonPathWriter::with_expand_dots(expand_dots_enabled);
+    for path in paths {
+        json_path.push(path);
     }
+    json_path.set_end();
+    let mut term = Term::with_type_and_field(Type::Json, json_field);
 
-    pub fn wrap(term_buffer: &'a mut Term, expand_dots_enabled: bool) -> Self {
-        term_buffer.clear_with_type(Type::Json);
-        let mut path_stack = Vec::with_capacity(10);
-        path_stack.push(0);
-        Self {
-            term_buffer,
-            path_stack,
-            expand_dots_enabled,
-        }
-    }
-
-    fn trim_to_end_of_path(&mut self) {
-        let end_of_path = *self.path_stack.last().unwrap();
-        self.term_buffer.truncate_value_bytes(end_of_path);
-    }
-
-    pub fn close_path_and_set_type(&mut self, typ: Type) {
-        self.trim_to_end_of_path();
-        self.term_buffer.set_json_path_end();
-        self.term_buffer.append_bytes(&[typ.to_code()]);
-    }
-
-    // TODO: Remove this function and use JsonPathWriter instead.
-    pub fn push_path_segment(&mut self, segment: &str) {
-        // the path stack should never be empty.
-        self.trim_to_end_of_path();
-
-        if self.path_stack.len() > 1 {
-            self.term_buffer.set_json_path_separator();
-        }
-        let appended_segment = self.term_buffer.append_bytes(segment.as_bytes());
-        if self.expand_dots_enabled {
-            // We need to replace `.` by JSON_PATH_SEGMENT_SEP.
-            replace_in_place(b'.', JSON_PATH_SEGMENT_SEP, appended_segment);
-        }
-        self.term_buffer.add_json_path_separator();
-        self.path_stack.push(self.term_buffer.len_bytes());
-    }
-
-    pub fn pop_path_segment(&mut self) {
-        self.path_stack.pop();
-        assert!(!self.path_stack.is_empty());
-        self.trim_to_end_of_path();
-    }
-
-    /// Returns the json path of the term being currently built.
-    #[cfg(test)]
-    pub(crate) fn path(&self) -> &[u8] {
-        let end_of_path = self.path_stack.last().cloned().unwrap_or(1);
-        &self.term().serialized_value_bytes()[..end_of_path - 1]
-    }
-
-    pub(crate) fn set_fast_value<T: FastValue>(&mut self, val: T) {
-        self.close_path_and_set_type(T::to_type());
-        let value = if T::to_type() == Type::Date {
-            DateTime::from_u64(val.to_u64())
-                .truncate(DATE_TIME_PRECISION_INDEXED)
-                .to_u64()
-        } else {
-            val.to_u64()
-        };
-        self.term_buffer
-            .append_bytes(value.to_be_bytes().as_slice());
-    }
-
-    pub fn set_str(&mut self, text: &str) {
-        self.close_path_and_set_type(Type::Str);
-        self.term_buffer.append_bytes(text.as_bytes());
-    }
-
-    pub fn term(&self) -> &Term {
-        self.term_buffer
-    }
+    term.append_bytes(json_path.as_str().as_bytes());
+    term
 }
 
 #[cfg(test)]
 mod tests {
-    use super::{split_json_path, JsonTermWriter};
-    use crate::schema::{Field, Type};
-    use crate::Term;
+    use super::split_json_path;
+    use crate::json_utils::term_from_json_paths;
+    use crate::schema::Field;
 
     #[test]
     fn test_json_writer() {
         let field = Field::from_field_id(1);
-        let mut term = Term::with_type_and_field(Type::Json, field);
-        let mut json_writer = JsonTermWriter::wrap(&mut term, false);
-        json_writer.push_path_segment("attributes");
-        json_writer.push_path_segment("color");
-        json_writer.set_str("red");
+
+        let mut term = term_from_json_paths(field, ["attributes", "color"].into_iter(), false);
+        term.append_type_and_str("red");
         assert_eq!(
-            format!("{:?}", json_writer.term()),
+            format!("{:?}", term),
             "Term(field=1, type=Json, path=attributes.color, type=Str, \"red\")"
         );
-        json_writer.set_str("blue");
-        assert_eq!(
-            format!("{:?}", json_writer.term()),
-            "Term(field=1, type=Json, path=attributes.color, type=Str, \"blue\")"
+
+        let mut term = term_from_json_paths(
+            field,
+            ["attributes", "dimensions", "width"].into_iter(),
+            false,
         );
-        json_writer.pop_path_segment();
-        json_writer.push_path_segment("dimensions");
-        json_writer.push_path_segment("width");
-        json_writer.set_fast_value(400i64);
+        term.append_type_and_fast_value(400i64);
         assert_eq!(
-            format!("{:?}", json_writer.term()),
+            format!("{:?}", term),
             "Term(field=1, type=Json, path=attributes.dimensions.width, type=I64, 400)"
         );
-        json_writer.pop_path_segment();
-        json_writer.push_path_segment("height");
-        json_writer.set_fast_value(300i64);
-        assert_eq!(
-            format!("{:?}", json_writer.term()),
-            "Term(field=1, type=Json, path=attributes.dimensions.height, type=I64, 300)"
-        );
     }
 
     #[test]
     fn test_string_term() {
         let field = Field::from_field_id(1);
-        let mut term = Term::with_type_and_field(Type::Json, field);
-        let mut json_writer = JsonTermWriter::wrap(&mut term, false);
-        json_writer.push_path_segment("color");
-        json_writer.set_str("red");
-        assert_eq!(
-            json_writer.term().serialized_term(),
-            b"\x00\x00\x00\x01jcolor\x00sred"
-        )
+        let mut term = term_from_json_paths(field, ["color"].into_iter(), false);
+        term.append_type_and_str("red");
+
+        assert_eq!(term.serialized_term(), b"\x00\x00\x00\x01jcolor\x00sred")
     }
 
     #[test]
     fn test_i64_term() {
         let field = Field::from_field_id(1);
-        let mut term = Term::with_type_and_field(Type::Json, field);
-        let mut json_writer = JsonTermWriter::wrap(&mut term, false);
-        json_writer.push_path_segment("color");
-        json_writer.set_fast_value(-4i64);
+        let mut term = term_from_json_paths(field, ["color"].into_iter(), false);
+        term.append_type_and_fast_value(-4i64);
+
         assert_eq!(
-            json_writer.term().serialized_term(),
-            b"\x00\x00\x00\x01jcolor\x00i\x7f\xff\xff\xff\xff\xff\xff\xfc"
+            term.value().as_serialized(),
+            b"jcolor\x00i\x7f\xff\xff\xff\xff\xff\xff\xfc"
         )
     }
 
     #[test]
     fn test_u64_term() {
         let field = Field::from_field_id(1);
-        let mut term = Term::with_type_and_field(Type::Json, field);
-        let mut json_writer = JsonTermWriter::wrap(&mut term, false);
-        json_writer.push_path_segment("color");
-        json_writer.set_fast_value(4u64);
+        let mut term = term_from_json_paths(field, ["color"].into_iter(), false);
+        term.append_type_and_fast_value(4u64);
+
         assert_eq!(
-            json_writer.term().serialized_term(),
-            b"\x00\x00\x00\x01jcolor\x00u\x00\x00\x00\x00\x00\x00\x00\x04"
+            term.value().as_serialized(),
+            b"jcolor\x00u\x00\x00\x00\x00\x00\x00\x00\x04"
         )
     }
 
     #[test]
     fn test_f64_term() {
         let field = Field::from_field_id(1);
-        let mut term = Term::with_type_and_field(Type::Json, field);
-        let mut json_writer = JsonTermWriter::wrap(&mut term, false);
-        json_writer.push_path_segment("color");
-        json_writer.set_fast_value(4.0f64);
+        let mut term = term_from_json_paths(field, ["color"].into_iter(), false);
+        term.append_type_and_fast_value(4.0f64);
         assert_eq!(
-            json_writer.term().serialized_term(),
-            b"\x00\x00\x00\x01jcolor\x00f\xc0\x10\x00\x00\x00\x00\x00\x00"
+            term.value().as_serialized(),
+            b"jcolor\x00f\xc0\x10\x00\x00\x00\x00\x00\x00"
         )
     }
 
     #[test]
     fn test_bool_term() {
         let field = Field::from_field_id(1);
-        let mut term = Term::with_type_and_field(Type::Json, field);
-        let mut json_writer = JsonTermWriter::wrap(&mut term, false);
-        json_writer.push_path_segment("color");
-        json_writer.set_fast_value(true);
+        let mut term = term_from_json_paths(field, ["color"].into_iter(), false);
+        term.append_type_and_fast_value(true);
         assert_eq!(
-            json_writer.term().serialized_term(),
-            b"\x00\x00\x00\x01jcolor\x00o\x00\x00\x00\x00\x00\x00\x00\x01"
+            term.value().as_serialized(),
+            b"jcolor\x00o\x00\x00\x00\x00\x00\x00\x00\x01"
         )
     }
 
-    #[test]
-    fn test_push_after_set_path_segment() {
-        let field = Field::from_field_id(1);
-        let mut term = Term::with_type_and_field(Type::Json, field);
-        let mut json_writer = JsonTermWriter::wrap(&mut term, false);
-        json_writer.push_path_segment("attribute");
-        json_writer.set_str("something");
-        json_writer.push_path_segment("color");
-        json_writer.set_str("red");
-        assert_eq!(
-            json_writer.term().serialized_term(),
-            b"\x00\x00\x00\x01jattribute\x01color\x00sred"
-        )
-    }
-
-    #[test]
-    fn test_pop_segment() {
-        let field = Field::from_field_id(1);
-        let mut term = Term::with_type_and_field(Type::Json, field);
-        let mut json_writer = JsonTermWriter::wrap(&mut term, false);
-        json_writer.push_path_segment("color");
-        json_writer.push_path_segment("hue");
-        json_writer.pop_path_segment();
-        json_writer.set_str("red");
-        assert_eq!(
-            json_writer.term().serialized_term(),
-            b"\x00\x00\x00\x01jcolor\x00sred"
-        )
-    }
-
-    #[test]
-    fn test_json_writer_path() {
-        let field = Field::from_field_id(1);
-        let mut term = Term::with_type_and_field(Type::Json, field);
-        let mut json_writer = JsonTermWriter::wrap(&mut term, false);
-        json_writer.push_path_segment("color");
-        assert_eq!(json_writer.path(), b"color");
-        json_writer.push_path_segment("hue");
-        assert_eq!(json_writer.path(), b"color\x01hue");
-        json_writer.set_str("pink");
-        assert_eq!(json_writer.path(), b"color\x01hue");
-    }
-
-    #[test]
-    fn test_json_path_expand_dots_disabled() {
-        let field = Field::from_field_id(1);
-        let mut term = Term::with_type_and_field(Type::Json, field);
-        let mut json_writer = JsonTermWriter::wrap(&mut term, false);
-        json_writer.push_path_segment("color.hue");
-        assert_eq!(json_writer.path(), b"color.hue");
-    }
-
-    #[test]
-    fn test_json_path_expand_dots_enabled() {
-        let field = Field::from_field_id(1);
-        let mut term = Term::with_type_and_field(Type::Json, field);
-        let mut json_writer = JsonTermWriter::wrap(&mut term, true);
-        json_writer.push_path_segment("color.hue");
-        assert_eq!(json_writer.path(), b"color\x01hue");
-    }
-
-    #[test]
-    fn test_json_path_expand_dots_enabled_pop_segment() {
-        let field = Field::from_field_id(1);
-        let mut term = Term::with_type_and_field(Type::Json, field);
-        let mut json_writer = JsonTermWriter::wrap(&mut term, true);
-        json_writer.push_path_segment("hello");
-        assert_eq!(json_writer.path(), b"hello");
-        json_writer.push_path_segment("color.hue");
-        assert_eq!(json_writer.path(), b"hello\x01color\x01hue");
-        json_writer.pop_path_segment();
-        assert_eq!(json_writer.path(), b"hello");
-    }
-
     #[test]
     fn test_split_json_path_simple() {
         let json_path = split_json_path("titi.toto");

src/core/tests.rs

@@ -1,9 +1,9 @@
 use crate::collector::Count;
 use crate::directory::{RamDirectory, WatchCallback};
 use crate::indexer::{LogMergePolicy, NoMergePolicy};
-use crate::json_utils::JsonTermWriter;
+use crate::json_utils::term_from_json_paths;
 use crate::query::TermQuery;
-use crate::schema::{Field, IndexRecordOption, Schema, Type, INDEXED, STRING, TEXT};
+use crate::schema::{Field, IndexRecordOption, Schema, INDEXED, STRING, TEXT};
 use crate::tokenizer::TokenizerManager;
 use crate::{
     Directory, DocSet, Index, IndexBuilder, IndexReader, IndexSettings, IndexWriter, Postings,
@@ -416,16 +416,12 @@ fn test_non_text_json_term_freq() {
     let searcher = reader.searcher();
     let segment_reader = searcher.segment_reader(0u32);
     let inv_idx = segment_reader.inverted_index(field).unwrap();
-    let mut term = Term::with_type_and_field(Type::Json, field);
-    let mut json_term_writer = JsonTermWriter::wrap(&mut term, false);
-    json_term_writer.push_path_segment("tenant_id");
-    json_term_writer.close_path_and_set_type(Type::U64);
-    json_term_writer.set_fast_value(75u64);
+
+    let mut term = term_from_json_paths(field, ["tenant_id"].iter().cloned(), false);
+    term.append_type_and_fast_value(75u64);
+
     let postings = inv_idx
-        .read_postings(
-            json_term_writer.term(),
-            IndexRecordOption::WithFreqsAndPositions,
-        )
+        .read_postings(&term, IndexRecordOption::WithFreqsAndPositions)
         .unwrap()
         .unwrap();
     assert_eq!(postings.doc(), 0);
@@ -454,16 +450,12 @@ fn test_non_text_json_term_freq_bitpacked() {
     let searcher = reader.searcher();
     let segment_reader = searcher.segment_reader(0u32);
     let inv_idx = segment_reader.inverted_index(field).unwrap();
-    let mut term = Term::with_type_and_field(Type::Json, field);
-    let mut json_term_writer = JsonTermWriter::wrap(&mut term, false);
-    json_term_writer.push_path_segment("tenant_id");
-    json_term_writer.close_path_and_set_type(Type::U64);
-    json_term_writer.set_fast_value(75u64);
+
+    let mut term = term_from_json_paths(field, ["tenant_id"].iter().cloned(), false);
+    term.append_type_and_fast_value(75u64);
+
     let mut postings = inv_idx
-        .read_postings(
-            json_term_writer.term(),
-            IndexRecordOption::WithFreqsAndPositions,
-        )
+        .read_postings(&term, IndexRecordOption::WithFreqsAndPositions)
         .unwrap()
         .unwrap();
     assert_eq!(postings.doc(), 0);

src/functional_test.rs

@@ -1,3 +1,5 @@
+#![allow(deprecated)] // Remove with index sorting
+
 use std::collections::HashSet;
 
 use rand::{thread_rng, Rng};

src/index/index.rs

@@ -20,7 +20,7 @@ use crate::indexer::segment_updater::save_metas;
 use crate::indexer::{IndexWriter, SingleSegmentIndexWriter};
 use crate::reader::{IndexReader, IndexReaderBuilder};
 use crate::schema::document::Document;
-use crate::schema::{Field, FieldType, Schema};
+use crate::schema::{Field, FieldType, Schema, Type};
 use crate::tokenizer::{TextAnalyzer, TokenizerManager};
 use crate::SegmentReader;
 
@@ -248,6 +248,15 @@ impl IndexBuilder {
                         sort_by_field.field
                     )));
                 }
+                let supported_field_types = [Type::I64, Type::U64, Type::F64, Type::Date];
+                let field_type = entry.field_type().value_type();
+                if !supported_field_types.contains(&field_type) {
+                    return Err(TantivyError::InvalidArgument(format!(
+                        "Unsupported field type in sort_by_field: {:?}. Supported field types: \
+                         {:?} ",
+                        field_type, supported_field_types,
+                    )));
+                }
             }
             Ok(())
         } else {

src/index/inverted_index_reader.rs

@@ -1,12 +1,13 @@
 use std::io;
 
+use common::json_path_writer::JSON_END_OF_PATH;
 use common::BinarySerializable;
 use fnv::FnvHashSet;
 
 use crate::directory::FileSlice;
 use crate::positions::PositionReader;
 use crate::postings::{BlockSegmentPostings, SegmentPostings, TermInfo};
-use crate::schema::{IndexRecordOption, Term, Type, JSON_END_OF_PATH};
+use crate::schema::{IndexRecordOption, Term, Type};
 use crate::termdict::TermDictionary;
 
 /// The inverted index reader is in charge of accessing

src/index/segment_reader.rs

@@ -406,7 +406,7 @@ impl SegmentReader {
     }
 
     /// Returns an iterator that will iterate over the alive document ids
-    pub fn doc_ids_alive(&self) -> Box<dyn Iterator<Item = DocId> + '_> {
+    pub fn doc_ids_alive(&self) -> Box<dyn Iterator<Item = DocId> + Send + '_> {
         if let Some(alive_bitset) = &self.alive_bitset_opt {
             Box::new(alive_bitset.iter_alive())
         } else {

src/indexer/doc_id_mapping.rs

@@ -159,7 +159,7 @@ mod tests_indexsorting {
     use crate::indexer::NoMergePolicy;
     use crate::query::QueryParser;
     use crate::schema::*;
-    use crate::{DocAddress, Index, IndexSettings, IndexSortByField, Order};
+    use crate::{DocAddress, Index, IndexBuilder, IndexSettings, IndexSortByField, Order};
 
     fn create_test_index(
         index_settings: Option<IndexSettings>,
@@ -557,4 +557,28 @@ mod tests_indexsorting {
             &[2000, 8000, 3000]
         );
     }
+
+    #[test]
+    fn test_text_sort() -> crate::Result<()> {
+        let mut schema_builder = SchemaBuilder::new();
+        schema_builder.add_text_field("id", STRING | FAST | STORED);
+        schema_builder.add_text_field("name", TEXT | STORED);
+
+        let resp = IndexBuilder::new()
+            .schema(schema_builder.build())
+            .settings(IndexSettings {
+                sort_by_field: Some(IndexSortByField {
+                    field: "id".to_string(),
+                    order: Order::Asc,
+                }),
+                ..Default::default()
+            })
+            .create_in_ram();
+        assert!(resp
+            .unwrap_err()
+            .to_string()
+            .contains("Unsupported field type"));
+
+        Ok(())
+    }
 }

src/indexer/mod.rs

@@ -156,16 +156,24 @@ mod tests_mmap {
     }
     #[test]
     fn test_json_field_1byte() {
-        // Test when field name contains a 1 byte, which has special meaning in tantivy.
+        // Test when field name contains a '1' byte, which has special meaning in tantivy.
+        // The 1 byte can be addressed as '1' byte or '.'.
         let field_name_in = "\u{0001}";
         let field_name_out = "\u{0001}";
         test_json_field_name(field_name_in, field_name_out);
 
-        // Test when field name contains a 1 byte, which has special meaning in tantivy.
+        // Test when field name contains a '1' byte, which has special meaning in tantivy.
         let field_name_in = "\u{0001}";
         let field_name_out = ".";
         test_json_field_name(field_name_in, field_name_out);
     }
+    #[test]
+    fn test_json_field_dot() {
+        // Test when field name contains a '.'
+        let field_name_in = ".";
+        let field_name_out = ".";
+        test_json_field_name(field_name_in, field_name_out);
+    }
     fn test_json_field_name(field_name_in: &str, field_name_out: &str) {
         let mut schema_builder = Schema::builder();
 
@@ -205,10 +213,10 @@ mod tests_mmap {
         let reader = index.reader().unwrap();
         let searcher = reader.searcher();
         let parse_query = QueryParser::for_index(&index, Vec::new());
-        let test_query = |field_name: &str| {
-            let query = parse_query.parse_query(field_name).unwrap();
+        let test_query = |query_str: &str| {
+            let query = parse_query.parse_query(query_str).unwrap();
             let num_docs = searcher.search(&query, &Count).unwrap();
-            assert_eq!(num_docs, 1);
+            assert_eq!(num_docs, 1, "{}", query_str);
         };
         test_query(format!("json.{field_name_out}:test1").as_str());
         test_query(format!("json.a{field_name_out}:test2").as_str());

src/indexer/segment_writer.rs

@@ -1,4 +1,3 @@
-use columnar::MonotonicallyMappableToU64;
 use common::JsonPathWriter;
 use itertools::Itertools;
 use tokenizer_api::BoxTokenStream;
@@ -15,7 +14,8 @@ use crate::postings::{
     PerFieldPostingsWriter, PostingsWriter,
 };
 use crate::schema::document::{Document, ReferenceValue, Value};
-use crate::schema::{FieldEntry, FieldType, Schema, Term, DATE_TIME_PRECISION_INDEXED};
+use crate::schema::indexing_term::IndexingTerm;
+use crate::schema::{FieldEntry, FieldType, Schema};
 use crate::store::{StoreReader, StoreWriter};
 use crate::tokenizer::{FacetTokenizer, PreTokenizedStream, TextAnalyzer, Tokenizer};
 use crate::{DocId, Opstamp, SegmentComponent, TantivyError};
@@ -70,7 +70,7 @@ pub struct SegmentWriter {
     pub(crate) json_path_writer: JsonPathWriter,
     pub(crate) doc_opstamps: Vec<Opstamp>,
     per_field_text_analyzers: Vec<TextAnalyzer>,
-    term_buffer: Term,
+    term_buffer: IndexingTerm,
     schema: Schema,
 }
 
@@ -126,7 +126,7 @@ impl SegmentWriter {
             )?,
             doc_opstamps: Vec::with_capacity(1_000),
             per_field_text_analyzers,
-            term_buffer: Term::with_capacity(16),
+            term_buffer: IndexingTerm::new(),
             schema,
         })
     }
@@ -195,7 +195,7 @@ impl SegmentWriter {
             let (term_buffer, ctx) = (&mut self.term_buffer, &mut self.ctx);
             let postings_writer: &mut dyn PostingsWriter =
                 self.per_field_postings_writers.get_for_field_mut(field);
-            term_buffer.clear_with_field_and_type(field_entry.field_type().value_type(), field);
+            term_buffer.clear_with_field(field);
 
             match field_entry.field_type() {
                 FieldType::Facet(_) => {
@@ -271,8 +271,7 @@ impl SegmentWriter {
 
                         num_vals += 1;
                         let date_val = value.as_datetime().ok_or_else(make_schema_error)?;
-                        term_buffer
-                            .set_u64(date_val.truncate(DATE_TIME_PRECISION_INDEXED).to_u64());
+                        term_buffer.set_date(date_val);
                         postings_writer.subscribe(doc_id, 0u32, term_buffer, ctx);
                     }
                     if field_entry.has_fieldnorms() {
@@ -332,7 +331,7 @@ impl SegmentWriter {
 
                         num_vals += 1;
                         let bytes = value.as_bytes().ok_or_else(make_schema_error)?;
-                        term_buffer.set_bytes(bytes);
+                        term_buffer.set_value_bytes(bytes);
                         postings_writer.subscribe(doc_id, 0u32, term_buffer, ctx);
                     }
                     if field_entry.has_fieldnorms() {
@@ -496,14 +495,14 @@ mod tests {
     use tempfile::TempDir;
 
     use crate::collector::{Count, TopDocs};
-    use crate::core::json_utils::JsonTermWriter;
     use crate::directory::RamDirectory;
+    use crate::fastfield::FastValue;
+    use crate::json_utils::term_from_json_paths;
     use crate::postings::TermInfo;
     use crate::query::{PhraseQuery, QueryParser};
     use crate::schema::document::Value;
     use crate::schema::{
-        Document, IndexRecordOption, Schema, TextFieldIndexing, TextOptions, Type, STORED, STRING,
-        TEXT,
+        Document, IndexRecordOption, Schema, TextFieldIndexing, TextOptions, STORED, STRING, TEXT,
     };
     use crate::store::{Compressor, StoreReader, StoreWriter};
     use crate::time::format_description::well_known::Rfc3339;
@@ -645,115 +644,117 @@ mod tests {
         let inv_idx = segment_reader.inverted_index(json_field).unwrap();
         let term_dict = inv_idx.terms();
 
-        let mut term = Term::with_type_and_field(Type::Json, json_field);
         let mut term_stream = term_dict.stream().unwrap();
 
-        let mut json_term_writer = JsonTermWriter::wrap(&mut term, false);
+        let term_from_path = |paths: &[&str]| -> Term {
+            term_from_json_paths(json_field, paths.iter().cloned(), false)
+        };
 
-        json_term_writer.push_path_segment("bool");
-        json_term_writer.set_fast_value(true);
+        fn set_fast_val<T: FastValue>(val: T, mut term: Term) -> Term {
+            term.append_type_and_fast_value(val);
+            term
+        }
+        fn set_str(val: &str, mut term: Term) -> Term {
+            term.append_type_and_str(val);
+            term
+        }
+
+        let term = term_from_path(&["bool"]);
         assert!(term_stream.advance());
         assert_eq!(
             term_stream.key(),
-            json_term_writer.term().serialized_value_bytes()
+            set_fast_val(true, term).serialized_value_bytes()
         );
 
-        json_term_writer.pop_path_segment();
-        json_term_writer.push_path_segment("complexobject");
-        json_term_writer.push_path_segment("field.with.dot");
-        json_term_writer.set_fast_value(1i64);
+        let term = term_from_path(&["complexobject", "field.with.dot"]);
+
         assert!(term_stream.advance());
         assert_eq!(
             term_stream.key(),
-            json_term_writer.term().serialized_value_bytes()
+            set_fast_val(1i64, term).serialized_value_bytes()
         );
 
-        json_term_writer.pop_path_segment();
-        json_term_writer.pop_path_segment();
-        json_term_writer.push_path_segment("date");
-        json_term_writer.set_fast_value(DateTime::from_utc(
-            OffsetDateTime::parse("1985-04-12T23:20:50.52Z", &Rfc3339).unwrap(),
-        ));
+        // Date
+        let term = term_from_path(&["date"]);
+
         assert!(term_stream.advance());
         assert_eq!(
             term_stream.key(),
-            json_term_writer.term().serialized_value_bytes()
+            set_fast_val(
+                DateTime::from_utc(
+                    OffsetDateTime::parse("1985-04-12T23:20:50.52Z", &Rfc3339).unwrap(),
+                ),
+                term
+            )
+            .serialized_value_bytes()
         );
 
-        json_term_writer.pop_path_segment();
-        json_term_writer.push_path_segment("float");
-        json_term_writer.set_fast_value(-0.2f64);
+        // Float
+        let term = term_from_path(&["float"]);
         assert!(term_stream.advance());
         assert_eq!(
             term_stream.key(),
-            json_term_writer.term().serialized_value_bytes()
+            set_fast_val(-0.2f64, term).serialized_value_bytes()
         );
 
-        json_term_writer.pop_path_segment();
-        json_term_writer.push_path_segment("my_arr");
-        json_term_writer.set_fast_value(2i64);
+        // Number In Array
+        let term = term_from_path(&["my_arr"]);
         assert!(term_stream.advance());
         assert_eq!(
             term_stream.key(),
-            json_term_writer.term().serialized_value_bytes()
+            set_fast_val(2i64, term).serialized_value_bytes()
         );
 
-        json_term_writer.set_fast_value(3i64);
+        let term = term_from_path(&["my_arr"]);
         assert!(term_stream.advance());
         assert_eq!(
             term_stream.key(),
-            json_term_writer.term().serialized_value_bytes()
+            set_fast_val(3i64, term).serialized_value_bytes()
         );
 
-        json_term_writer.set_fast_value(4i64);
+        let term = term_from_path(&["my_arr"]);
         assert!(term_stream.advance());
         assert_eq!(
             term_stream.key(),
-            json_term_writer.term().serialized_value_bytes()
+            set_fast_val(4i64, term).serialized_value_bytes()
         );
 
-        json_term_writer.push_path_segment("my_key");
-        json_term_writer.set_str("tokens");
+        // El in Array
+        let term = term_from_path(&["my_arr", "my_key"]);
         assert!(term_stream.advance());
         assert_eq!(
             term_stream.key(),
-            json_term_writer.term().serialized_value_bytes()
+            set_str("tokens", term).serialized_value_bytes()
         );
-
-        json_term_writer.set_str("two");
+        let term = term_from_path(&["my_arr", "my_key"]);
         assert!(term_stream.advance());
         assert_eq!(
             term_stream.key(),
-            json_term_writer.term().serialized_value_bytes()
+            set_str("two", term).serialized_value_bytes()
         );
 
-        json_term_writer.pop_path_segment();
-        json_term_writer.pop_path_segment();
-        json_term_writer.push_path_segment("signed");
-        json_term_writer.set_fast_value(-2i64);
+        // Signed
+        let term = term_from_path(&["signed"]);
         assert!(term_stream.advance());
         assert_eq!(
             term_stream.key(),
-            json_term_writer.term().serialized_value_bytes()
+            set_fast_val(-2i64, term).serialized_value_bytes()
         );
 
-        json_term_writer.pop_path_segment();
-        json_term_writer.push_path_segment("toto");
-        json_term_writer.set_str("titi");
+        let term = term_from_path(&["toto"]);
         assert!(term_stream.advance());
         assert_eq!(
             term_stream.key(),
-            json_term_writer.term().serialized_value_bytes()
+            set_str("titi", term).serialized_value_bytes()
         );
-
-        json_term_writer.pop_path_segment();
-        json_term_writer.push_path_segment("unsigned");
-        json_term_writer.set_fast_value(1i64);
+        // Unsigned
+        let term = term_from_path(&["unsigned"]);
         assert!(term_stream.advance());
         assert_eq!(
             term_stream.key(),
-            json_term_writer.term().serialized_value_bytes()
+            set_fast_val(1i64, term).serialized_value_bytes()
         );
+
         assert!(!term_stream.advance());
     }
 
@@ -774,14 +775,9 @@ mod tests {
         let searcher = reader.searcher();
         let segment_reader = searcher.segment_reader(0u32);
         let inv_index = segment_reader.inverted_index(json_field).unwrap();
-        let mut term = Term::with_type_and_field(Type::Json, json_field);
-        let mut json_term_writer = JsonTermWriter::wrap(&mut term, false);
-        json_term_writer.push_path_segment("mykey");
-        json_term_writer.set_str("token");
-        let term_info = inv_index
-            .get_term_info(json_term_writer.term())
-            .unwrap()
-            .unwrap();
+        let mut term = term_from_json_paths(json_field, ["mykey"].into_iter(), false);
+        term.append_type_and_str("token");
+        let term_info = inv_index.get_term_info(&term).unwrap().unwrap();
         assert_eq!(
             term_info,
             TermInfo {
@@ -818,14 +814,9 @@ mod tests {
         let searcher = reader.searcher();
         let segment_reader = searcher.segment_reader(0u32);
         let inv_index = segment_reader.inverted_index(json_field).unwrap();
-        let mut term = Term::with_type_and_field(Type::Json, json_field);
-        let mut json_term_writer = JsonTermWriter::wrap(&mut term, false);
-        json_term_writer.push_path_segment("mykey");
-        json_term_writer.set_str("two tokens");
-        let term_info = inv_index
-            .get_term_info(json_term_writer.term())
-            .unwrap()
-            .unwrap();
+        let mut term = term_from_json_paths(json_field, ["mykey"].into_iter(), false);
+        term.append_type_and_str("two tokens");
+        let term_info = inv_index.get_term_info(&term).unwrap().unwrap();
         assert_eq!(
             term_info,
             TermInfo {
@@ -863,16 +854,18 @@ mod tests {
         writer.commit().unwrap();
         let reader = index.reader().unwrap();
         let searcher = reader.searcher();
-        let mut term = Term::with_type_and_field(Type::Json, json_field);
-        let mut json_term_writer = JsonTermWriter::wrap(&mut term, false);
-        json_term_writer.push_path_segment("mykey");
-        json_term_writer.push_path_segment("field");
-        json_term_writer.set_str("hello");
-        let hello_term = json_term_writer.term().clone();
-        json_term_writer.set_str("nothello");
-        let nothello_term = json_term_writer.term().clone();
-        json_term_writer.set_str("happy");
-        let happy_term = json_term_writer.term().clone();
+
+        let term = term_from_json_paths(json_field, ["mykey", "field"].into_iter(), false);
+
+        let mut hello_term = term.clone();
+        hello_term.append_type_and_str("hello");
+
+        let mut nothello_term = term.clone();
+        nothello_term.append_type_and_str("nothello");
+
+        let mut happy_term = term.clone();
+        happy_term.append_type_and_str("happy");
+
         let phrase_query = PhraseQuery::new(vec![hello_term, happy_term.clone()]);
         assert_eq!(searcher.search(&phrase_query, &Count).unwrap(), 1);
         let phrase_query = PhraseQuery::new(vec![nothello_term, happy_term]);

src/lib.rs

@@ -255,7 +255,7 @@ pub struct Version {
 
 impl fmt::Debug for Version {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
-        write!(f, "{}", self.to_string())
+        fmt::Display::fmt(self, f)
     }
 }
 
@@ -266,9 +266,10 @@ static VERSION: Lazy<Version> = Lazy::new(|| Version {
     index_format_version: INDEX_FORMAT_VERSION,
 });
 
-impl ToString for Version {
-    fn to_string(&self) -> String {
-        format!(
+impl fmt::Display for Version {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        write!(
+            f,
             "tantivy v{}.{}.{}, index_format v{}",
             self.major, self.minor, self.patch, self.index_format_version
         )

src/postings/json_postings_writer.rs

@@ -1,5 +1,6 @@
 use std::io;
 
+use common::json_path_writer::JSON_END_OF_PATH;
 use stacker::Addr;
 
 use crate::indexer::doc_id_mapping::DocIdMapping;
@@ -7,9 +8,10 @@ use crate::indexer::path_to_unordered_id::OrderedPathId;
 use crate::postings::postings_writer::SpecializedPostingsWriter;
 use crate::postings::recorder::{BufferLender, DocIdRecorder, Recorder};
 use crate::postings::{FieldSerializer, IndexingContext, IndexingPosition, PostingsWriter};
-use crate::schema::{Field, Type, JSON_END_OF_PATH};
+use crate::schema::indexing_term::IndexingTerm;
+use crate::schema::{Field, Type, ValueBytes};
 use crate::tokenizer::TokenStream;
-use crate::{DocId, Term};
+use crate::DocId;
 
 /// The `JsonPostingsWriter` is odd in that it relies on a hidden contract:
 ///
@@ -33,7 +35,7 @@ impl<Rec: Recorder> PostingsWriter for JsonPostingsWriter<Rec> {
         &mut self,
         doc: crate::DocId,
         pos: u32,
-        term: &crate::Term,
+        term: &IndexingTerm,
         ctx: &mut IndexingContext,
     ) {
         self.non_str_posting_writer.subscribe(doc, pos, term, ctx);
@@ -43,7 +45,7 @@ impl<Rec: Recorder> PostingsWriter for JsonPostingsWriter<Rec> {
         &mut self,
         doc_id: DocId,
         token_stream: &mut dyn TokenStream,
-        term_buffer: &mut Term,
+        term_buffer: &mut IndexingTerm,
         ctx: &mut IndexingContext,
         indexing_position: &mut IndexingPosition,
     ) {
@@ -65,42 +67,40 @@ impl<Rec: Recorder> PostingsWriter for JsonPostingsWriter<Rec> {
         ctx: &IndexingContext,
         serializer: &mut FieldSerializer,
     ) -> io::Result<()> {
-        let mut term_buffer = Term::with_capacity(48);
+        let mut term_buffer = JsonTermSerializer(Vec::with_capacity(48));
         let mut buffer_lender = BufferLender::default();
-        term_buffer.clear_with_field_and_type(Type::Json, Field::from_field_id(0));
         let mut prev_term_id = u32::MAX;
         let mut term_path_len = 0; // this will be set in the first iteration
         for (_field, path_id, term, addr) in term_addrs {
             if prev_term_id != path_id.path_id() {
-                term_buffer.truncate_value_bytes(0);
+                term_buffer.clear();
                 term_buffer.append_path(ordered_id_to_path[path_id.path_id() as usize].as_bytes());
                 term_buffer.append_bytes(&[JSON_END_OF_PATH]);
-                term_path_len = term_buffer.len_bytes();
+                term_path_len = term_buffer.len();
                 prev_term_id = path_id.path_id();
             }
-            term_buffer.truncate_value_bytes(term_path_len);
+            term_buffer.truncate(term_path_len);
             term_buffer.append_bytes(term);
-            if let Some(json_value) = term_buffer.value().as_json_value_bytes() {
-                let typ = json_value.typ();
-                if typ == Type::Str {
-                    SpecializedPostingsWriter::<Rec>::serialize_one_term(
-                        term_buffer.serialized_value_bytes(),
-                        *addr,
-                        doc_id_map,
-                        &mut buffer_lender,
-                        ctx,
-                        serializer,
-                    )?;
-                } else {
-                    SpecializedPostingsWriter::<DocIdRecorder>::serialize_one_term(
-                        term_buffer.serialized_value_bytes(),
-                        *addr,
-                        doc_id_map,
-                        &mut buffer_lender,
-                        ctx,
-                        serializer,
-                    )?;
-                }
+            let json_value = ValueBytes::wrap(term);
+            let typ = json_value.typ();
+            if typ == Type::Str {
+                SpecializedPostingsWriter::<Rec>::serialize_one_term(
+                    term_buffer.as_bytes(),
+                    *addr,
+                    doc_id_map,
+                    &mut buffer_lender,
+                    ctx,
+                    serializer,
+                )?;
+            } else {
+                SpecializedPostingsWriter::<DocIdRecorder>::serialize_one_term(
+                    term_buffer.as_bytes(),
+                    *addr,
+                    doc_id_map,
+                    &mut buffer_lender,
+                    ctx,
+                    serializer,
+                )?;
             }
         }
         Ok(())
@@ -110,3 +110,40 @@ impl<Rec: Recorder> PostingsWriter for JsonPostingsWriter<Rec> {
         self.str_posting_writer.total_num_tokens() + self.non_str_posting_writer.total_num_tokens()
     }
 }
+
+struct JsonTermSerializer(Vec<u8>);
+impl JsonTermSerializer {
+    #[inline]
+    pub fn append_path(&mut self, bytes: &[u8]) {
+        if bytes.contains(&0u8) {
+            self.0
+                .extend(bytes.iter().map(|&b| if b == 0 { b'0' } else { b }));
+        } else {
+            self.0.extend_from_slice(bytes);
+        }
+    }
+
+    /// Appends value bytes to the Term.
+    ///
+    /// This function returns the segment that has just been added.
+    #[inline]
+    pub fn append_bytes(&mut self, bytes: &[u8]) -> &mut [u8] {
+        let len_before = self.0.len();
+        self.0.extend_from_slice(bytes);
+        &mut self.0[len_before..]
+    }
+
+    fn clear(&mut self) {
+        self.0.clear();
+    }
+    fn truncate(&mut self, len: usize) {
+        self.0.truncate(len);
+    }
+    fn len(&self) -> usize {
+        self.0.len()
+    }
+
+    fn as_bytes(&self) -> &[u8] {
+        &self.0
+    }
+}

src/postings/postings_writer.rs

@@ -11,7 +11,8 @@ use crate::postings::recorder::{BufferLender, Recorder};
 use crate::postings::{
     FieldSerializer, IndexingContext, InvertedIndexSerializer, PerFieldPostingsWriter,
 };
-use crate::schema::{Field, Schema, Term, Type};
+use crate::schema::indexing_term::{get_field_from_indexing_term, IndexingTerm};
+use crate::schema::{Field, Schema, Type};
 use crate::tokenizer::{Token, TokenStream, MAX_TOKEN_LEN};
 use crate::DocId;
 
@@ -60,14 +61,14 @@ pub(crate) fn serialize_postings(
     let mut term_offsets: Vec<(Field, OrderedPathId, &[u8], Addr)> =
         Vec::with_capacity(ctx.term_index.len());
     term_offsets.extend(ctx.term_index.iter().map(|(key, addr)| {
-        let field = Term::wrap(key).field();
+        let field = get_field_from_indexing_term(key);
         if schema.get_field_entry(field).field_type().value_type() == Type::Json {
-            let byte_range_path = 5..5 + 4;
+            let byte_range_path = 4..4 + 4;
             let unordered_id = u32::from_be_bytes(key[byte_range_path.clone()].try_into().unwrap());
             let path_id = unordered_id_to_ordered_id[unordered_id as usize];
             (field, path_id, &key[byte_range_path.end..], addr)
         } else {
-            (field, 0.into(), &key[5..], addr)
+            (field, 0.into(), &key[4..], addr)
         }
     }));
     // Sort by field, path, and term
@@ -114,7 +115,7 @@ 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: &IndexingTerm, ctx: &mut IndexingContext);
 
     /// Serializes the postings on disk.
     /// The actual serialization format is handled by the `PostingsSerializer`.
@@ -132,7 +133,7 @@ pub(crate) trait PostingsWriter: Send + Sync {
         &mut self,
         doc_id: DocId,
         token_stream: &mut dyn TokenStream,
-        term_buffer: &mut Term,
+        term_buffer: &mut IndexingTerm,
         ctx: &mut IndexingContext,
         indexing_position: &mut IndexingPosition,
     ) {
@@ -203,26 +204,35 @@ impl<Rec: Recorder> SpecializedPostingsWriter<Rec> {
 
 impl<Rec: Recorder> PostingsWriter for SpecializedPostingsWriter<Rec> {
     #[inline]
-    fn subscribe(&mut self, doc: DocId, position: u32, term: &Term, ctx: &mut IndexingContext) {
-        debug_assert!(term.serialized_term().len() >= 4);
+    fn subscribe(
+        &mut self,
+        doc: DocId,
+        position: u32,
+        term: &IndexingTerm,
+        ctx: &mut IndexingContext,
+    ) {
+        debug_assert!(term.serialized_for_hashmap().len() >= 4);
         self.total_num_tokens += 1;
         let (term_index, arena) = (&mut ctx.term_index, &mut ctx.arena);
-        term_index.mutate_or_create(term.serialized_term(), |opt_recorder: Option<Rec>| {
-            if let Some(mut recorder) = opt_recorder {
-                let current_doc = recorder.current_doc();
-                if current_doc != doc {
-                    recorder.close_doc(arena);
+        term_index.mutate_or_create(
+            term.serialized_for_hashmap(),
+            |opt_recorder: Option<Rec>| {
+                if let Some(mut recorder) = opt_recorder {
+                    let current_doc = recorder.current_doc();
+                    if current_doc != doc {
+                        recorder.close_doc(arena);
+                        recorder.new_doc(doc, arena);
+                    }
+                    recorder.record_position(position, arena);
+                    recorder
+                } else {
+                    let mut recorder = Rec::default();
                     recorder.new_doc(doc, arena);
+                    recorder.record_position(position, arena);
+                    recorder
                 }
-                recorder.record_position(position, arena);
-                recorder
-            } else {
-                let mut recorder = Rec::default();
-                recorder.new_doc(doc, arena);
-                recorder.record_position(position, arena);
-                recorder
-            }
-        });
+            },
+        );
     }
 
     fn serialize(

src/query/fuzzy_query.rs

@@ -3,7 +3,7 @@ use once_cell::sync::OnceCell;
 use tantivy_fst::Automaton;
 
 use crate::query::{AutomatonWeight, EnableScoring, Query, Weight};
-use crate::schema::{Term, Type};
+use crate::schema::Term;
 use crate::TantivyError::InvalidArgument;
 
 pub(crate) struct DfaWrapper(pub DFA);
@@ -133,40 +133,33 @@ impl FuzzyTermQuery {
 
         let term_value = self.term.value();
 
-        let term_text = if term_value.typ() == Type::Json {
-            if let Some(json_path_type) = term_value.json_path_type() {
-                if json_path_type != Type::Str {
-                    return Err(InvalidArgument(format!(
-                        "The fuzzy term query requires a string path type for a json term. Found \
-                         {:?}",
-                        json_path_type
-                    )));
-                }
+        let get_automaton = |term_text: &str| {
+            if self.prefix {
+                automaton_builder.build_prefix_dfa(term_text)
+            } else {
+                automaton_builder.build_dfa(term_text)
             }
-
-            std::str::from_utf8(self.term.serialized_value_bytes()).map_err(|_| {
-                InvalidArgument(
-                    "Failed to convert json term value bytes to utf8 string.".to_string(),
-                )
-            })?
-        } else {
-            term_value.as_str().ok_or_else(|| {
-                InvalidArgument("The fuzzy term query requires a string term.".to_string())
-            })?
-        };
-        let automaton = if self.prefix {
-            automaton_builder.build_prefix_dfa(term_text)
-        } else {
-            automaton_builder.build_dfa(term_text)
         };
 
-        if let Some((json_path_bytes, _)) = term_value.as_json() {
+        if let Some((json_path_bytes, _term_value)) = term_value.as_json() {
+            let term_text =
+                std::str::from_utf8(self.term.serialized_value_bytes()).map_err(|_| {
+                    InvalidArgument(
+                        "Failed to convert json term value bytes to utf8 string.".to_string(),
+                    )
+                })?;
+
+            let automaton = get_automaton(term_text);
             Ok(AutomatonWeight::new_for_json_path(
                 self.term.field(),
                 DfaWrapper(automaton),
                 json_path_bytes,
             ))
         } else {
+            let term_text = term_value.as_str().ok_or_else(|| {
+                InvalidArgument("The fuzzy term query requires a string term.".to_string())
+            })?;
+            let automaton = get_automaton(term_text);
             Ok(AutomatonWeight::new(
                 self.term.field(),
                 DfaWrapper(automaton),

src/query/phrase_prefix_query/phrase_prefix_query.rs

@@ -137,7 +137,7 @@ impl Query for PhrasePrefixQuery {
             // There are no prefix. Let's just match the suffix.
             let end_term =
                 if let Some(end_value) = prefix_end(self.prefix.1.serialized_value_bytes()) {
-                    let mut end_term = Term::with_capacity(end_value.len());
+                    let mut end_term = Term::new();
                     end_term.set_field_and_type(self.field, self.prefix.1.typ());
                     end_term.append_bytes(&end_value);
                     Bound::Excluded(end_term)

src/query/query_parser/query_parser.rs

@@ -10,10 +10,10 @@ use query_grammar::{UserInputAst, UserInputBound, UserInputLeaf, UserInputLitera
 use rustc_hash::FxHashMap;
 
 use super::logical_ast::*;
-use crate::core::json_utils::{
-    convert_to_fast_value_and_get_term, set_string_and_get_terms, JsonTermWriter,
-};
 use crate::index::Index;
+use crate::json_utils::{
+    convert_to_fast_value_and_append_to_json_term, split_json_path, term_from_json_paths,
+};
 use crate::query::range_query::{is_type_valid_for_fastfield_range_query, RangeQuery};
 use crate::query::{
     AllQuery, BooleanQuery, BoostQuery, EmptyQuery, FuzzyTermQuery, Occur, PhrasePrefixQuery,
@@ -965,20 +965,33 @@ fn generate_literals_for_json_object(
         })?;
     let index_record_option = text_options.index_option();
     let mut logical_literals = Vec::new();
-    let mut term = Term::with_capacity(100);
-    let mut json_term_writer = JsonTermWriter::from_field_and_json_path(
-        field,
-        json_path,
-        json_options.is_expand_dots_enabled(),
-        &mut term,
-    );
-    if let Some(term) = convert_to_fast_value_and_get_term(&mut json_term_writer, phrase) {
+
+    let paths = split_json_path(json_path);
+    let get_term_with_path = || {
+        term_from_json_paths(
+            field,
+            paths.iter().map(|el| el.as_str()),
+            json_options.is_expand_dots_enabled(),
+        )
+    };
+
+    // Try to convert the phrase to a fast value
+    if let Some(term) = convert_to_fast_value_and_append_to_json_term(get_term_with_path(), phrase)
+    {
         logical_literals.push(LogicalLiteral::Term(term));
     }
-    let terms = set_string_and_get_terms(&mut json_term_writer, phrase, &mut text_analyzer);
-    drop(json_term_writer);
-    if terms.len() <= 1 {
-        for (_, term) in terms {
+
+    // Try to tokenize the phrase and create Terms.
+    let mut positions_and_terms = Vec::<(usize, Term)>::new();
+    let mut token_stream = text_analyzer.token_stream(phrase);
+    token_stream.process(&mut |token| {
+        let mut term = get_term_with_path();
+        term.append_type_and_str(&token.text);
+        positions_and_terms.push((token.position, term.clone()));
+    });
+
+    if positions_and_terms.len() <= 1 {
+        for (_, term) in positions_and_terms {
             logical_literals.push(LogicalLiteral::Term(term));
         }
         return Ok(logical_literals);
@@ -989,7 +1002,7 @@ fn generate_literals_for_json_object(
         ));
     }
     logical_literals.push(LogicalLiteral::Phrase {
-        terms,
+        terms: positions_and_terms,
         slop: 0,
         prefix: false,
     });

src/schema/indexing_term.rs

@@ -0,0 +1,147 @@
+use std::net::Ipv6Addr;
+
+use columnar::{MonotonicallyMappableToU128, MonotonicallyMappableToU64};
+
+use super::date_time_options::DATE_TIME_PRECISION_INDEXED;
+use super::Field;
+use crate::fastfield::FastValue;
+use crate::schema::Type;
+use crate::DateTime;
+
+/// IndexingTerm represents is the serialized information of a term during indexing.
+/// It's a serialized representation over different types.
+///
+/// It actually wraps a `Vec<u8>`.
+///
+/// The format is as follow:
+/// `[field id: u32][serialized value]`
+///
+/// For JSON it equals to:
+/// `[field id: u32][path id: u32][type code: u8][serialized value]`
+///
+/// The format is chosen to easily partition the terms by field during serialization, as all terms
+/// are stored in one hashmap.
+#[derive(Clone)]
+pub(crate) struct IndexingTerm(Vec<u8>);
+
+/// The number of bytes used as for the field id by `Term`.
+const FIELD_ID_LENGTH: usize = 4;
+
+impl IndexingTerm {
+    /// Create a new IndexingTerm.
+    pub fn new() -> IndexingTerm {
+        let mut data = Vec::with_capacity(FIELD_ID_LENGTH + 32);
+        data.resize(FIELD_ID_LENGTH, 0u8);
+        IndexingTerm(data)
+    }
+
+    /// Is empty if there are no value bytes.
+    pub fn is_empty(&self) -> bool {
+        self.0.len() == FIELD_ID_LENGTH
+    }
+
+    /// Removes the value_bytes and set the field
+    pub(crate) fn clear_with_field(&mut self, field: Field) {
+        self.truncate_value_bytes(0);
+        self.0[0..4].clone_from_slice(field.field_id().to_be_bytes().as_ref());
+    }
+
+    /// Sets a u64 value in the term.
+    ///
+    /// U64 are serialized using (8-byte) BigEndian
+    /// representation.
+    /// The use of BigEndian has the benefit of preserving
+    /// the natural order of the values.
+    pub fn set_u64(&mut self, val: u64) {
+        self.set_fast_value(val);
+    }
+
+    /// Sets a `DateTime` value in the term.
+    pub fn set_date(&mut self, val: DateTime) {
+        self.set_fast_value(val);
+    }
+
+    /// Sets a `i64` value in the term.
+    pub fn set_i64(&mut self, val: i64) {
+        self.set_fast_value(val);
+    }
+
+    /// Sets a `f64` value in the term.
+    pub fn set_f64(&mut self, val: f64) {
+        self.set_fast_value(val);
+    }
+
+    /// Sets a `bool` value in the term.
+    pub fn set_bool(&mut self, val: bool) {
+        self.set_fast_value(val);
+    }
+
+    fn set_fast_value<T: FastValue>(&mut self, val: T) {
+        self.truncate_value_bytes(0);
+        self.append_fast_value(val);
+    }
+
+    /// Sets a `Ipv6Addr` value in the term.
+    pub fn set_ip_addr(&mut self, val: Ipv6Addr) {
+        self.set_value_bytes(val.to_u128().to_be_bytes().as_ref());
+    }
+
+    /// Sets the value bytes of the term.
+    pub fn set_value_bytes(&mut self, bytes: &[u8]) {
+        self.truncate_value_bytes(0);
+        self.0.extend(bytes);
+    }
+
+    /// Append a type marker + fast value to a term.
+    /// This is used in JSON type to append a fast value after the path.
+    ///
+    /// It will not clear existing bytes.
+    pub(crate) fn append_type_and_fast_value<T: FastValue>(&mut self, val: T) {
+        self.0.push(T::to_type().to_code());
+        self.append_fast_value(val)
+    }
+
+    /// Append a fast value to a term.
+    ///
+    /// It will not clear existing bytes.
+    pub fn append_fast_value<T: FastValue>(&mut self, val: T) {
+        let value = if T::to_type() == Type::Date {
+            DateTime::from_u64(val.to_u64())
+                .truncate(DATE_TIME_PRECISION_INDEXED)
+                .to_u64()
+        } else {
+            val.to_u64()
+        };
+        self.0.extend(value.to_be_bytes().as_ref());
+    }
+
+    /// Truncates the value bytes of the term. Value and field type stays the same.
+    pub fn truncate_value_bytes(&mut self, len: usize) {
+        self.0.truncate(len + FIELD_ID_LENGTH);
+    }
+
+    /// The length of the bytes.
+    pub fn len_bytes(&self) -> usize {
+        self.0.len() - FIELD_ID_LENGTH
+    }
+
+    /// Appends bytes to the Term.
+    ///
+    /// This function returns the segment that has just been added.
+    #[inline]
+    pub fn append_bytes(&mut self, bytes: &[u8]) {
+        self.0.extend_from_slice(bytes);
+    }
+
+    /// Returns the serialized representation of Term.
+    /// This includes field_id, value bytes
+    #[inline]
+    pub fn serialized_for_hashmap(&self) -> &[u8] {
+        self.0.as_ref()
+    }
+}
+
+pub fn get_field_from_indexing_term(bytes: &[u8]) -> Field {
+    let field_id_bytes: [u8; 4] = bytes[..4].try_into().unwrap();
+    Field::from_field_id(u32::from_be_bytes(field_id_bytes))
+}

src/schema/mod.rs

@@ -109,6 +109,7 @@
 pub mod document;
 mod facet;
 mod facet_options;
+pub(crate) mod indexing_term;
 mod schema;
 pub(crate) mod term;
 
@@ -146,7 +147,7 @@ pub use self::json_object_options::JsonObjectOptions;
 pub use self::named_field_document::NamedFieldDocument;
 pub use self::numeric_options::NumericOptions;
 pub use self::schema::{Schema, SchemaBuilder};
-pub use self::term::{Term, ValueBytes, JSON_END_OF_PATH};
+pub use self::term::{Term, ValueBytes};
 pub use self::text_options::{TextFieldIndexing, TextOptions, STRING, TEXT};
 
 /// Validator for a potential `field_name`.

src/schema/term.rs

@@ -1,8 +1,9 @@
-use std::hash::{Hash, Hasher};
+use std::hash::Hash;
 use std::net::Ipv6Addr;
 use std::{fmt, str};
 
 use columnar::{MonotonicallyMappableToU128, MonotonicallyMappableToU64};
+use common::json_path_writer::{JSON_END_OF_PATH, JSON_PATH_SEGMENT_SEP_STR};
 
 use super::date_time_options::DATE_TIME_PRECISION_INDEXED;
 use super::Field;
@@ -10,15 +11,6 @@ use crate::fastfield::FastValue;
 use crate::schema::{Facet, Type};
 use crate::DateTime;
 
-/// Separates the different segments of a json path.
-pub const JSON_PATH_SEGMENT_SEP: u8 = 1u8;
-pub const JSON_PATH_SEGMENT_SEP_STR: &str =
-    unsafe { std::str::from_utf8_unchecked(&[JSON_PATH_SEGMENT_SEP]) };
-
-/// Separates the json path and the value in
-/// a JSON term binary representation.
-pub const JSON_END_OF_PATH: u8 = 0u8;
-
 /// Term represents the value that the token can take.
 /// It's a serialized representation over different types.
 ///
@@ -26,37 +18,45 @@ pub const JSON_END_OF_PATH: u8 = 0u8;
 /// 4 bytes are the field id, and the last byte is the type.
 ///
 /// The serialized value `ValueBytes` is considered everything after the 4 first bytes (term id).
-#[derive(Clone)]
-pub struct Term<B = Vec<u8>>(B)
-where B: AsRef<[u8]>;
+#[derive(Clone, Hash, PartialEq, Ord, PartialOrd, Eq)]
+pub struct Term(Vec<u8>);
+impl Default for Term {
+    fn default() -> Self {
+        Self::new()
+    }
+}
 
 /// The number of bytes used as metadata by `Term`.
 const TERM_METADATA_LENGTH: usize = 5;
 
 impl Term {
-    /// Create a new Term with a buffer with a given capacity.
-    pub fn with_capacity(capacity: usize) -> Term {
-        let mut data = Vec::with_capacity(TERM_METADATA_LENGTH + capacity);
+    /// Create a new Term
+    pub fn new() -> Term {
+        let mut data = Vec::with_capacity(TERM_METADATA_LENGTH + 32);
         data.resize(TERM_METADATA_LENGTH, 0u8);
         Term(data)
     }
 
     pub(crate) fn with_type_and_field(typ: Type, field: Field) -> Term {
-        let mut term = Self::with_capacity(8);
-        term.set_field_and_type(field, typ);
-        term
+        Self::with_bytes_and_field_and_payload(typ, field, &[])
     }
 
     fn with_bytes_and_field_and_payload(typ: Type, field: Field, bytes: &[u8]) -> Term {
-        let mut term = Self::with_capacity(bytes.len());
+        let mut term = Self::new();
         term.set_field_and_type(field, typ);
         term.0.extend_from_slice(bytes);
         term
     }
 
+    /// Sets a fast value in the term.
+    ///
+    /// fast values are converted to u64 and then serialized using (8-byte) BigEndian
+    /// representation.
+    /// The use of BigEndian has the benefit of preserving
+    /// the natural order of the values.
     fn from_fast_value<T: FastValue>(field: Field, val: &T) -> Term {
         let mut term = Self::with_type_and_field(T::to_type(), field);
-        term.set_u64(val.to_u64());
+        term.set_bytes(val.to_u64().to_be_bytes().as_ref());
         term
     }
 
@@ -78,7 +78,7 @@ impl Term {
     /// Builds a term given a field, and a `Ipv6Addr`-value
     pub fn from_field_ip_addr(field: Field, ip_addr: Ipv6Addr) -> Term {
         let mut term = Self::with_type_and_field(Type::IpAddr, field);
-        term.set_ip_addr(ip_addr);
+        term.set_bytes(ip_addr.to_u128().to_be_bytes().as_ref());
         term
     }
 
@@ -123,52 +123,16 @@ impl Term {
         Term::with_bytes_and_field_and_payload(Type::Bytes, field, bytes)
     }
 
-    /// Removes the value_bytes and set the field and type code.
-    pub(crate) fn clear_with_field_and_type(&mut self, typ: Type, field: Field) {
-        self.truncate_value_bytes(0);
-        self.set_field_and_type(field, typ);
-    }
-
     /// Removes the value_bytes and set the type code.
     pub fn clear_with_type(&mut self, typ: Type) {
         self.truncate_value_bytes(0);
         self.0[4] = typ.to_code();
     }
 
-    /// Sets a u64 value in the term.
+    /// Append a type marker + fast value to a term.
+    /// This is used in JSON type to append a fast value after the path.
     ///
-    /// U64 are serialized using (8-byte) BigEndian
-    /// representation.
-    /// The use of BigEndian has the benefit of preserving
-    /// the natural order of the values.
-    pub fn set_u64(&mut self, val: u64) {
-        self.set_fast_value(val);
-    }
-
-    /// Sets a `i64` value in the term.
-    pub fn set_i64(&mut self, val: i64) {
-        self.set_fast_value(val);
-    }
-
-    /// Sets a `DateTime` value in the term.
-    pub fn set_date(&mut self, date: DateTime) {
-        self.set_fast_value(date);
-    }
-
-    /// Sets a `f64` value in the term.
-    pub fn set_f64(&mut self, val: f64) {
-        self.set_fast_value(val);
-    }
-
-    /// Sets a `bool` value in the term.
-    pub fn set_bool(&mut self, val: bool) {
-        self.set_fast_value(val);
-    }
-
-    fn set_fast_value<T: FastValue>(&mut self, val: T) {
-        self.set_bytes(val.to_u64().to_be_bytes().as_ref());
-    }
-
+    /// It will not clear existing bytes.
     pub(crate) fn append_type_and_fast_value<T: FastValue>(&mut self, val: T) {
         self.0.push(T::to_type().to_code());
         let value = if T::to_type() == Type::Date {
@@ -181,32 +145,26 @@ impl Term {
         self.0.extend(value.to_be_bytes().as_ref());
     }
 
-    /// Sets a `Ipv6Addr` value in the term.
-    pub fn set_ip_addr(&mut self, val: Ipv6Addr) {
-        self.set_bytes(val.to_u128().to_be_bytes().as_ref());
+    /// Append a string type marker + string to a term.
+    /// This is used in JSON type to append a str after the path.
+    ///
+    /// It will not clear existing bytes.
+    pub(crate) fn append_type_and_str(&mut self, val: &str) {
+        self.0.push(Type::Str.to_code());
+        self.0.extend(val.as_bytes().as_ref());
     }
 
     /// Sets the value of a `Bytes` field.
-    pub fn set_bytes(&mut self, bytes: &[u8]) {
+    fn set_bytes(&mut self, bytes: &[u8]) {
         self.truncate_value_bytes(0);
         self.0.extend(bytes);
     }
 
-    /// Set the texts only, keeping the field untouched.
-    pub fn set_text(&mut self, text: &str) {
-        self.set_bytes(text.as_bytes());
-    }
-
     /// Truncates the value bytes of the term. Value and field type stays the same.
     pub fn truncate_value_bytes(&mut self, len: usize) {
         self.0.truncate(len + TERM_METADATA_LENGTH);
     }
 
-    /// The length of the bytes.
-    pub fn len_bytes(&self) -> usize {
-        self.0.len() - TERM_METADATA_LENGTH
-    }
-
     /// Appends value bytes to the Term.
     ///
     /// This function returns the segment that has just been added.
@@ -217,78 +175,19 @@ impl Term {
         &mut self.0[len_before..]
     }
 
-    /// Appends json path bytes to the Term.
-    /// If the path contains 0 bytes, they are replaced by a "0" string.
-    /// The 0 byte is used to mark the end of the path.
-    ///
-    /// This function returns the segment that has just been added.
-    #[inline]
-    pub fn append_path(&mut self, bytes: &[u8]) -> &mut [u8] {
-        let len_before = self.0.len();
-        if bytes.contains(&0u8) {
-            self.0
-                .extend(bytes.iter().map(|&b| if b == 0 { b'0' } else { b }));
-        } else {
-            self.0.extend_from_slice(bytes);
-        }
-        &mut self.0[len_before..]
-    }
-
-    /// Appends a JSON_PATH_SEGMENT_SEP to the term.
-    /// Only used for JSON type.
-    #[inline]
-    pub fn add_json_path_separator(&mut self) {
-        self.0.push(JSON_PATH_SEGMENT_SEP);
-    }
-    /// Sets the current end to JSON_END_OF_PATH.
-    /// Only used for JSON type.
-    #[inline]
-    pub fn set_json_path_end(&mut self) {
-        let buffer_len = self.0.len();
-        self.0[buffer_len - 1] = JSON_END_OF_PATH;
-    }
-    /// Sets the current end to JSON_PATH_SEGMENT_SEP.
-    /// Only used for JSON type.
-    #[inline]
-    pub fn set_json_path_separator(&mut self) {
-        let buffer_len = self.0.len();
-        self.0[buffer_len - 1] = JSON_PATH_SEGMENT_SEP;
-    }
-}
-
-impl<B> Term<B>
-where B: AsRef<[u8]>
-{
-    /// Wraps a object holding bytes
-    pub fn wrap(data: B) -> Term<B> {
-        Term(data)
-    }
-
     /// Return the type of the term.
     pub fn typ(&self) -> Type {
         self.value().typ()
     }
 
-    /// Returns the field.
-    pub fn field(&self) -> Field {
-        let field_id_bytes: [u8; 4] = (&self.0.as_ref()[..4]).try_into().unwrap();
-        Field::from_field_id(u32::from_be_bytes(field_id_bytes))
-    }
-
     /// Returns the serialized representation of the value.
     /// (this does neither include the field id nor the value type.)
     ///
     /// If the term is a string, its value is utf-8 encoded.
     /// If the term is a u64, its value is encoded according
     /// to `byteorder::BigEndian`.
-    pub fn serialized_value_bytes(&self) -> &[u8] {
-        &self.0.as_ref()[TERM_METADATA_LENGTH..]
-    }
-
-    /// Returns the value of the term.
-    /// address or JSON path + value. (this does not include the field.)
-    pub fn value(&self) -> ValueBytes<&[u8]> {
-        ValueBytes::wrap(&self.0.as_ref()[4..])
+    pub(crate) fn serialized_value_bytes(&self) -> &[u8] {
+        &self.0[TERM_METADATA_LENGTH..]
     }
 
     /// Returns the serialized representation of Term.
@@ -297,9 +196,22 @@ where B: AsRef<[u8]>
     /// Do NOT rely on this byte representation in the index.
     /// This value is likely to change in the future.
     #[inline]
+    #[cfg(test)]
     pub fn serialized_term(&self) -> &[u8] {
         self.0.as_ref()
     }
+
+    /// Returns the field.
+    pub fn field(&self) -> Field {
+        let field_id_bytes: [u8; 4] = (&self.0[..4]).try_into().unwrap();
+        Field::from_field_id(u32::from_be_bytes(field_id_bytes))
+    }
+
+    /// Returns the value of the term.
+    /// address or JSON path + value. (this does not include the field.)
+    pub fn value(&self) -> ValueBytes<&[u8]> {
+        ValueBytes::wrap(&self.0[4..])
+    }
 }
 
 /// ValueBytes represents a serialized value.
@@ -330,18 +242,10 @@ where B: AsRef<[u8]>
     }
 
     /// Return the type of the term.
-    pub fn typ(&self) -> Type {
+    pub(crate) fn typ(&self) -> Type {
         Type::from_code(self.typ_code()).expect("The term has an invalid type code")
     }
 
-    /// Returns the `u64` value stored in a term.
-    ///
-    /// Returns `None` if the term is not of the u64 type, or if the term byte representation
-    /// is invalid.
-    pub fn as_u64(&self) -> Option<u64> {
-        self.get_fast_type::<u64>()
-    }
-
     fn get_fast_type<T: FastValue>(&self) -> Option<T> {
         if self.typ() != T::to_type() {
             return None;
@@ -351,38 +255,6 @@ where B: AsRef<[u8]>
         Some(T::from_u64(value_u64))
     }
 
-    /// Returns the `i64` value stored in a term.
-    ///
-    /// Returns `None` if the term is not of the i64 type, or if the term byte representation
-    /// is invalid.
-    pub fn as_i64(&self) -> Option<i64> {
-        self.get_fast_type::<i64>()
-    }
-
-    /// Returns the `f64` value stored in a term.
-    ///
-    /// Returns `None` if the term is not of the f64 type, or if the term byte representation
-    /// is invalid.
-    pub fn as_f64(&self) -> Option<f64> {
-        self.get_fast_type::<f64>()
-    }
-
-    /// Returns the `bool` value stored in a term.
-    ///
-    /// Returns `None` if the term is not of the bool type, or if the term byte representation
-    /// is invalid.
-    pub fn as_bool(&self) -> Option<bool> {
-        self.get_fast_type::<bool>()
-    }
-
-    /// Returns the `Date` value stored in a term.
-    ///
-    /// Returns `None` if the term is not of the Date type, or if the term byte representation
-    /// is invalid.
-    pub fn as_date(&self) -> Option<DateTime> {
-        self.get_fast_type::<DateTime>()
-    }
-
     /// Returns the text associated with the term.
     ///
     /// Returns `None` if the field is not of string type
@@ -398,7 +270,7 @@ where B: AsRef<[u8]>
     ///
     /// Returns `None` if the field is not of facet type
     /// or if the bytes are not valid utf-8.
-    pub fn as_facet(&self) -> Option<Facet> {
+    pub(crate) fn as_facet(&self) -> Option<Facet> {
         if self.typ() != Type::Facet {
             return None;
         }
@@ -409,7 +281,7 @@ where B: AsRef<[u8]>
     /// Returns the bytes associated with the term.
     ///
     /// Returns `None` if the field is not of bytes type.
-    pub fn as_bytes(&self) -> Option<&[u8]> {
+    pub(crate) fn as_bytes(&self) -> Option<&[u8]> {
         if self.typ() != Type::Bytes {
             return None;
         }
@@ -417,7 +289,7 @@ where B: AsRef<[u8]>
     }
 
     /// Returns a `Ipv6Addr` value from the term.
-    pub fn as_ip_addr(&self) -> Option<Ipv6Addr> {
+    pub(crate) fn as_ip_addr(&self) -> Option<Ipv6Addr> {
         if self.typ() != Type::IpAddr {
             return None;
         }
@@ -425,15 +297,6 @@ where B: AsRef<[u8]>
         Some(Ipv6Addr::from_u128(ip_u128))
     }
 
-    /// Returns the json path type.
-    ///
-    /// Returns `None` if the value is not JSON.
-    pub fn json_path_type(&self) -> Option<Type> {
-        let json_value_bytes = self.as_json_value_bytes()?;
-
-        Some(json_value_bytes.typ())
-    }
-
     /// Returns the json path bytes (including the JSON_END_OF_PATH byte),
     /// and the encoded ValueBytes after the json path.
     ///
@@ -450,18 +313,6 @@ where B: AsRef<[u8]>
         Some((json_path_bytes, ValueBytes::wrap(term)))
     }
 
-    /// Returns the encoded ValueBytes after the json path.
-    ///
-    /// Returns `None` if the value is not JSON.
-    pub(crate) fn as_json_value_bytes(&self) -> Option<ValueBytes<&[u8]>> {
-        if self.typ() != Type::Json {
-            return None;
-        }
-        let bytes = self.value_bytes();
-        let pos = bytes.iter().cloned().position(|b| b == JSON_END_OF_PATH)?;
-        Some(ValueBytes::wrap(&bytes[pos + 1..]))
-    }
-
     /// Returns the serialized value of ValueBytes without the type.
     fn value_bytes(&self) -> &[u8] {
         &self.0.as_ref()[1..]
@@ -484,20 +335,20 @@ where B: AsRef<[u8]>
                 write_opt(f, s)?;
             }
             Type::U64 => {
-                write_opt(f, self.as_u64())?;
+                write_opt(f, self.get_fast_type::<u64>())?;
             }
             Type::I64 => {
-                write_opt(f, self.as_i64())?;
+                write_opt(f, self.get_fast_type::<i64>())?;
             }
             Type::F64 => {
-                write_opt(f, self.as_f64())?;
+                write_opt(f, self.get_fast_type::<f64>())?;
             }
             Type::Bool => {
-                write_opt(f, self.as_bool())?;
+                write_opt(f, self.get_fast_type::<bool>())?;
             }
             // TODO pretty print these types too.
             Type::Date => {
-                write_opt(f, self.as_date())?;
+                write_opt(f, self.get_fast_type::<DateTime>())?;
             }
             Type::Facet => {
                 write_opt(f, self.as_facet())?;
@@ -523,40 +374,6 @@ where B: AsRef<[u8]>
     }
 }
 
-impl<B> Ord for Term<B>
-where B: AsRef<[u8]>
-{
-    fn cmp(&self, other: &Self) -> std::cmp::Ordering {
-        self.serialized_term().cmp(other.serialized_term())
-    }
-}
-
-impl<B> PartialOrd for Term<B>
-where B: AsRef<[u8]>
-{
-    fn partial_cmp(&self, other: &Self) -> Option<std::cmp::Ordering> {
-        Some(self.cmp(other))
-    }
-}
-
-impl<B> PartialEq for Term<B>
-where B: AsRef<[u8]>
-{
-    fn eq(&self, other: &Self) -> bool {
-        self.serialized_term() == other.serialized_term()
-    }
-}
-
-impl<B> Eq for Term<B> where B: AsRef<[u8]> {}
-
-impl<B> Hash for Term<B>
-where B: AsRef<[u8]>
-{
-    fn hash<H: Hasher>(&self, state: &mut H) {
-        self.0.as_ref().hash(state)
-    }
-}
-
 fn write_opt<T: std::fmt::Debug>(f: &mut fmt::Formatter, val_opt: Option<T>) -> fmt::Result {
     if let Some(val) = val_opt {
         write!(f, "{val:?}")?;
@@ -564,13 +381,11 @@ fn write_opt<T: std::fmt::Debug>(f: &mut fmt::Formatter, val_opt: Option<T>) ->
     Ok(())
 }
 
-impl<B> fmt::Debug for Term<B>
-where B: AsRef<[u8]>
-{
+impl fmt::Debug for Term {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
         let field_id = self.field().field_id();
         write!(f, "Term(field={field_id}, ")?;
-        let value_bytes = ValueBytes::wrap(&self.0.as_ref()[4..]);
+        let value_bytes = ValueBytes::wrap(&self.0[4..]);
         value_bytes.debug_value_bytes(f)?;
         write!(f, ")",)?;
         Ok(())
@@ -592,38 +407,4 @@ mod tests {
         assert_eq!(term.typ(), Type::Str);
         assert_eq!(term.value().as_str(), Some("test"))
     }
-
-    /// Size (in bytes) of the buffer of a fast value (u64, i64, f64, or date) term.
-    /// <field> + <type byte> + <value len>
-    ///
-    /// - <field> is a big endian encoded u32 field id
-    /// - <type_byte>'s most significant bit expresses whether the term is a json term or not
-    /// The remaining 7 bits are used to encode the type of the value.
-    /// If this is a JSON term, the type is the type of the leaf of the json.
-    ///
-    /// - <value> is,  if this is not the json term, a binary representation specific to the type.
-    /// If it is a JSON Term, then it is prepended with the path that leads to this leaf value.
-    const FAST_VALUE_TERM_LEN: usize = 4 + 1 + 8;
-
-    #[test]
-    pub fn test_term_u64() {
-        let mut schema_builder = Schema::builder();
-        let count_field = schema_builder.add_u64_field("count", INDEXED);
-        let term = Term::from_field_u64(count_field, 983u64);
-        assert_eq!(term.field(), count_field);
-        assert_eq!(term.typ(), Type::U64);
-        assert_eq!(term.serialized_term().len(), FAST_VALUE_TERM_LEN);
-        assert_eq!(term.value().as_u64(), Some(983u64))
-    }
-
-    #[test]
-    pub fn test_term_bool() {
-        let mut schema_builder = Schema::builder();
-        let bool_field = schema_builder.add_bool_field("bool", INDEXED);
-        let term = Term::from_field_bool(bool_field, true);
-        assert_eq!(term.field(), bool_field);
-        assert_eq!(term.typ(), Type::Bool);
-        assert_eq!(term.serialized_term().len(), FAST_VALUE_TERM_LEN);
-        assert_eq!(term.value().as_bool(), Some(true))
-    }
 }

src/store/reader.rs

@@ -14,7 +14,7 @@ use super::Decompressor;
 use crate::directory::FileSlice;
 use crate::error::DataCorruption;
 use crate::fastfield::AliveBitSet;
-use crate::schema::document::{BinaryDocumentDeserializer, Document, DocumentDeserialize};
+use crate::schema::document::{BinaryDocumentDeserializer, DocumentDeserialize};
 use crate::space_usage::StoreSpaceUsage;
 use crate::store::index::Checkpoint;
 use crate::DocId;
@@ -235,7 +235,7 @@ impl StoreReader {
     /// Iterator over all Documents in their order as they are stored in the doc store.
     /// Use this, if you want to extract all Documents from the doc store.
     /// The `alive_bitset` has to be forwarded from the `SegmentReader` or the results may be wrong.
-    pub fn iter<'a: 'b, 'b, D: Document + DocumentDeserialize>(
+    pub fn iter<'a: 'b, 'b, D: DocumentDeserialize>(
         &'b self,
         alive_bitset: Option<&'a AliveBitSet>,
     ) -> impl Iterator<Item = crate::Result<D>> + 'b {

sstable/src/lib.rs

@@ -1,6 +1,5 @@
 use std::io::{self, Write};
 use std::ops::Range;
-use std::usize;
 
 use merge::ValueMerger;