chore: bump version to 0.14.1 (#6006)

* feat: remove own greatest fn (#5994)

* fix: prune primary key with multiple columns may use default value as statistics (#5996)

* test: incorrect test result when filtering pk with multiple columns

* fix: prune non first tag correctly

Distinguish no column and no stats and only use default value when no
column

* test: update test result

* refactor: rename test file

* test: add test for null filter

* fix: use StatValues for null counts

* test: drop table

* test: fix unstable flow test

* fix: check if memtable is empty by stats (#5989)

fix/checking-memtable-empty-and-stats:
 - **Refactor timestamp updates**: Simplified timestamp range updates in `PartitionTreeMemtable` and `TimeSeriesMemtable` by replacing `update_timestamp_range` with `fetch_max` and `fetch_min` methods for `max_timestamp` and `min_timestamp`.
   - Affected files: `partition_tree.rs`, `time_series.rs`

 - **Remove unused code**: Deleted the `update_timestamp_range` method from `WriteMetrics` and removed unnecessary imports.
   - Affected file: `stats.rs`

 - **Optimize memtable filtering**: Streamlined the check for empty memtables in `ScanRegion` by directly using `time_range`.
   - Affected file: `scan_region.rs`

* chore: bump version to 0.14.1

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>

---------

Signed-off-by: Zhenchi <zhongzc_arch@outlook.com>
Co-authored-by: dennis zhuang <killme2008@gmail.com>
Co-authored-by: Yingwen <realevenyag@gmail.com>
Co-authored-by: Lei, HUANG <6406592+v0y4g3r@users.noreply.github.com>

Cargo.lock

@@ -185,7 +185,7 @@ checksum = "d301b3b94cb4b2f23d7917810addbbaff90738e0ca2be692bd027e70d7e0330c"
 
 [[package]]
 name = "api"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "common-base",
  "common-decimal",
@@ -915,7 +915,7 @@ dependencies = [
 
 [[package]]
 name = "auth"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "async-trait",
@@ -1537,7 +1537,7 @@ dependencies = [
 
 [[package]]
 name = "cache"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "catalog",
  "common-error",
@@ -1561,7 +1561,7 @@ checksum = "37b2a672a2cb129a2e41c10b1224bb368f9f37a2b16b612598138befd7b37eb5"
 
 [[package]]
 name = "catalog"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "arrow 54.2.1",
@@ -1874,7 +1874,7 @@ checksum = "1462739cb27611015575c0c11df5df7601141071f07518d56fcc1be504cbec97"
 
 [[package]]
 name = "cli"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "async-trait",
  "auth",
@@ -1917,7 +1917,7 @@ dependencies = [
  "session",
  "snafu 0.8.5",
  "store-api",
- "substrait 0.14.0",
+ "substrait 0.14.1",
  "table",
  "tempfile",
  "tokio",
@@ -1926,7 +1926,7 @@ dependencies = [
 
 [[package]]
 name = "client"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "arc-swap",
@@ -1955,7 +1955,7 @@ dependencies = [
  "rand 0.9.0",
  "serde_json",
  "snafu 0.8.5",
- "substrait 0.14.0",
+ "substrait 0.14.1",
  "substrait 0.37.3",
  "tokio",
  "tokio-stream",
@@ -1996,7 +1996,7 @@ dependencies = [
 
 [[package]]
 name = "cmd"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "async-trait",
  "auth",
@@ -2056,7 +2056,7 @@ dependencies = [
  "similar-asserts",
  "snafu 0.8.5",
  "store-api",
- "substrait 0.14.0",
+ "substrait 0.14.1",
  "table",
  "temp-env",
  "tempfile",
@@ -2102,7 +2102,7 @@ checksum = "55b672471b4e9f9e95499ea597ff64941a309b2cdbffcc46f2cc5e2d971fd335"
 
 [[package]]
 name = "common-base"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "anymap2",
  "async-trait",
@@ -2124,11 +2124,11 @@ dependencies = [
 
 [[package]]
 name = "common-catalog"
-version = "0.14.0"
+version = "0.14.1"
 
 [[package]]
 name = "common-config"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "common-base",
  "common-error",
@@ -2153,7 +2153,7 @@ dependencies = [
 
 [[package]]
 name = "common-datasource"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "arrow 54.2.1",
  "arrow-schema 54.3.1",
@@ -2190,7 +2190,7 @@ dependencies = [
 
 [[package]]
 name = "common-decimal"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "bigdecimal 0.4.8",
  "common-error",
@@ -2203,7 +2203,7 @@ dependencies = [
 
 [[package]]
 name = "common-error"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "common-macro",
  "http 1.1.0",
@@ -2214,7 +2214,7 @@ dependencies = [
 
 [[package]]
 name = "common-frontend"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "async-trait",
  "common-error",
@@ -2224,7 +2224,7 @@ dependencies = [
 
 [[package]]
 name = "common-function"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "ahash 0.8.11",
  "api",
@@ -2277,7 +2277,7 @@ dependencies = [
 
 [[package]]
 name = "common-greptimedb-telemetry"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "async-trait",
  "common-runtime",
@@ -2294,7 +2294,7 @@ dependencies = [
 
 [[package]]
 name = "common-grpc"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "arrow-flight",
@@ -2325,7 +2325,7 @@ dependencies = [
 
 [[package]]
 name = "common-grpc-expr"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "common-base",
@@ -2344,7 +2344,7 @@ dependencies = [
 
 [[package]]
 name = "common-macro"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "arc-swap",
  "common-query",
@@ -2358,7 +2358,7 @@ dependencies = [
 
 [[package]]
 name = "common-mem-prof"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "common-error",
  "common-macro",
@@ -2371,7 +2371,7 @@ dependencies = [
 
 [[package]]
 name = "common-meta"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "anymap2",
  "api",
@@ -2432,7 +2432,7 @@ dependencies = [
 
 [[package]]
 name = "common-options"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "common-grpc",
  "humantime-serde",
@@ -2441,11 +2441,11 @@ dependencies = [
 
 [[package]]
 name = "common-plugins"
-version = "0.14.0"
+version = "0.14.1"
 
 [[package]]
 name = "common-pprof"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "common-error",
  "common-macro",
@@ -2457,7 +2457,7 @@ dependencies = [
 
 [[package]]
 name = "common-procedure"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "async-stream",
  "async-trait",
@@ -2484,7 +2484,7 @@ dependencies = [
 
 [[package]]
 name = "common-procedure-test"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "async-trait",
  "common-procedure",
@@ -2493,7 +2493,7 @@ dependencies = [
 
 [[package]]
 name = "common-query"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "async-trait",
@@ -2519,7 +2519,7 @@ dependencies = [
 
 [[package]]
 name = "common-recordbatch"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "arc-swap",
  "common-error",
@@ -2539,7 +2539,7 @@ dependencies = [
 
 [[package]]
 name = "common-runtime"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "async-trait",
  "clap 4.5.19",
@@ -2569,14 +2569,14 @@ dependencies = [
 
 [[package]]
 name = "common-session"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "strum 0.27.1",
 ]
 
 [[package]]
 name = "common-telemetry"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "atty",
  "backtrace",
@@ -2604,7 +2604,7 @@ dependencies = [
 
 [[package]]
 name = "common-test-util"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "client",
  "common-query",
@@ -2616,7 +2616,7 @@ dependencies = [
 
 [[package]]
 name = "common-time"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "arrow 54.2.1",
  "chrono",
@@ -2634,7 +2634,7 @@ dependencies = [
 
 [[package]]
 name = "common-version"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "build-data",
  "const_format",
@@ -2644,7 +2644,7 @@ dependencies = [
 
 [[package]]
 name = "common-wal"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "common-base",
  "common-error",
@@ -3572,7 +3572,7 @@ dependencies = [
 
 [[package]]
 name = "datanode"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "arrow-flight",
@@ -3624,7 +3624,7 @@ dependencies = [
  "session",
  "snafu 0.8.5",
  "store-api",
- "substrait 0.14.0",
+ "substrait 0.14.1",
  "table",
  "tokio",
  "toml 0.8.19",
@@ -3633,7 +3633,7 @@ dependencies = [
 
 [[package]]
 name = "datatypes"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "arrow 54.2.1",
  "arrow-array 54.2.1",
@@ -4259,7 +4259,7 @@ dependencies = [
 
 [[package]]
 name = "file-engine"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "async-trait",
@@ -4382,7 +4382,7 @@ checksum = "8bf7cc16383c4b8d58b9905a8509f02926ce3058053c056376248d958c9df1e8"
 
 [[package]]
 name = "flow"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "arrow 54.2.1",
@@ -4444,7 +4444,7 @@ dependencies = [
  "snafu 0.8.5",
  "store-api",
  "strum 0.27.1",
- "substrait 0.14.0",
+ "substrait 0.14.1",
  "table",
  "tokio",
  "tonic 0.12.3",
@@ -4499,7 +4499,7 @@ checksum = "6c2141d6d6c8512188a7891b4b01590a45f6dac67afb4f255c4124dbb86d4eaa"
 
 [[package]]
 name = "frontend"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "arc-swap",
@@ -4556,7 +4556,7 @@ dependencies = [
  "sqlparser 0.54.0 (git+https://github.com/GreptimeTeam/sqlparser-rs.git?rev=0cf6c04490d59435ee965edd2078e8855bd8471e)",
  "store-api",
  "strfmt",
- "substrait 0.14.0",
+ "substrait 0.14.1",
  "table",
  "tokio",
  "toml 0.8.19",
@@ -5795,7 +5795,7 @@ dependencies = [
 
 [[package]]
 name = "index"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "async-trait",
  "asynchronous-codec",
@@ -6605,7 +6605,7 @@ checksum = "a7a70ba024b9dc04c27ea2f0c0548feb474ec5c54bba33a7f72f873a39d07b24"
 
 [[package]]
 name = "log-query"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "chrono",
  "common-error",
@@ -6617,7 +6617,7 @@ dependencies = [
 
 [[package]]
 name = "log-store"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "async-stream",
  "async-trait",
@@ -6911,7 +6911,7 @@ dependencies = [
 
 [[package]]
 name = "meta-client"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "async-trait",
@@ -6939,7 +6939,7 @@ dependencies = [
 
 [[package]]
 name = "meta-srv"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "async-trait",
@@ -7029,7 +7029,7 @@ dependencies = [
 
 [[package]]
 name = "metric-engine"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "aquamarine",
@@ -7118,7 +7118,7 @@ dependencies = [
 
 [[package]]
 name = "mito2"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "aquamarine",
@@ -7824,7 +7824,7 @@ dependencies = [
 
 [[package]]
 name = "object-store"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "anyhow",
  "bytes",
@@ -8119,7 +8119,7 @@ dependencies = [
 
 [[package]]
 name = "operator"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "ahash 0.8.11",
  "api",
@@ -8168,7 +8168,7 @@ dependencies = [
  "sql",
  "sqlparser 0.54.0 (git+https://github.com/GreptimeTeam/sqlparser-rs.git?rev=0cf6c04490d59435ee965edd2078e8855bd8471e)",
  "store-api",
- "substrait 0.14.0",
+ "substrait 0.14.1",
  "table",
  "tokio",
  "tokio-util",
@@ -8423,7 +8423,7 @@ dependencies = [
 
 [[package]]
 name = "partition"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "async-trait",
@@ -8705,7 +8705,7 @@ checksum = "8b870d8c151b6f2fb93e84a13146138f05d02ed11c7e7c54f8826aaaf7c9f184"
 
 [[package]]
 name = "pipeline"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "ahash 0.8.11",
  "api",
@@ -8847,7 +8847,7 @@ dependencies = [
 
 [[package]]
 name = "plugins"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "auth",
  "clap 4.5.19",
@@ -9127,7 +9127,7 @@ dependencies = [
 
 [[package]]
 name = "promql"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "ahash 0.8.11",
  "async-trait",
@@ -9373,7 +9373,7 @@ dependencies = [
 
 [[package]]
 name = "puffin"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "async-compression 0.4.13",
  "async-trait",
@@ -9414,7 +9414,7 @@ dependencies = [
 
 [[package]]
 name = "query"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "ahash 0.8.11",
  "api",
@@ -9480,7 +9480,7 @@ dependencies = [
  "sqlparser 0.54.0 (git+https://github.com/GreptimeTeam/sqlparser-rs.git?rev=0cf6c04490d59435ee965edd2078e8855bd8471e)",
  "statrs",
  "store-api",
- "substrait 0.14.0",
+ "substrait 0.14.1",
  "table",
  "tokio",
  "tokio-stream",
@@ -10830,7 +10830,7 @@ dependencies = [
 
 [[package]]
 name = "servers"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "ahash 0.8.11",
  "api",
@@ -10950,7 +10950,7 @@ dependencies = [
 
 [[package]]
 name = "session"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "arc-swap",
@@ -11275,7 +11275,7 @@ dependencies = [
 
 [[package]]
 name = "sql"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "chrono",
@@ -11330,7 +11330,7 @@ dependencies = [
 
 [[package]]
 name = "sqlness-runner"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "async-trait",
  "clap 4.5.19",
@@ -11649,7 +11649,7 @@ dependencies = [
 
 [[package]]
 name = "store-api"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "aquamarine",
@@ -11798,7 +11798,7 @@ dependencies = [
 
 [[package]]
 name = "substrait"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "async-trait",
  "bytes",
@@ -11978,7 +11978,7 @@ dependencies = [
 
 [[package]]
 name = "table"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "async-trait",
@@ -12229,7 +12229,7 @@ checksum = "3369f5ac52d5eb6ab48c6b4ffdc8efbcad6b89c765749064ba298f2c68a16a76"
 
 [[package]]
 name = "tests-fuzz"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "arbitrary",
  "async-trait",
@@ -12273,7 +12273,7 @@ dependencies = [
 
 [[package]]
 name = "tests-integration"
-version = "0.14.0"
+version = "0.14.1"
 dependencies = [
  "api",
  "arrow-flight",
@@ -12340,7 +12340,7 @@ dependencies = [
  "sql",
  "sqlx",
  "store-api",
- "substrait 0.14.0",
+ "substrait 0.14.1",
  "table",
  "tempfile",
  "time",

Cargo.toml

@@ -68,7 +68,7 @@ members = [
 resolver = "2"
 
 [workspace.package]
-version = "0.14.0"
+version = "0.14.1"
 edition = "2021"
 license = "Apache-2.0"
 

src/common/function/src/scalars/timestamp.rs

@@ -13,10 +13,8 @@
 // limitations under the License.
 
 use std::sync::Arc;
-mod greatest;
 mod to_unixtime;
 
-use greatest::GreatestFunction;
 use to_unixtime::ToUnixtimeFunction;
 
 use crate::function_registry::FunctionRegistry;
@@ -26,6 +24,5 @@ pub(crate) struct TimestampFunction;
 impl TimestampFunction {
     pub fn register(registry: &FunctionRegistry) {
         registry.register(Arc::new(ToUnixtimeFunction));
-        registry.register(Arc::new(GreatestFunction));
     }
 }

src/common/function/src/scalars/timestamp/greatest.rs

@@ -1,328 +0,0 @@
-// Copyright 2023 Greptime Team
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use std::fmt::{self};
-
-use common_query::error::{
-    self, ArrowComputeSnafu, InvalidFuncArgsSnafu, Result, UnsupportedInputDataTypeSnafu,
-};
-use common_query::prelude::{Signature, Volatility};
-use datafusion::arrow::compute::kernels::cmp::gt;
-use datatypes::arrow::array::AsArray;
-use datatypes::arrow::compute::cast;
-use datatypes::arrow::compute::kernels::zip;
-use datatypes::arrow::datatypes::{
-    DataType as ArrowDataType, Date32Type, TimeUnit, TimestampMicrosecondType,
-    TimestampMillisecondType, TimestampNanosecondType, TimestampSecondType,
-};
-use datatypes::prelude::ConcreteDataType;
-use datatypes::types::TimestampType;
-use datatypes::vectors::{Helper, VectorRef};
-use snafu::{ensure, ResultExt};
-
-use crate::function::{Function, FunctionContext};
-
-#[derive(Clone, Debug, Default)]
-pub struct GreatestFunction;
-
-const NAME: &str = "greatest";
-
-macro_rules! gt_time_types {
-    ($ty: ident, $columns:expr) => {{
-        let column1 = $columns[0].to_arrow_array();
-        let column2 = $columns[1].to_arrow_array();
-
-        let column1 = column1.as_primitive::<$ty>();
-        let column2 = column2.as_primitive::<$ty>();
-        let boolean_array = gt(&column1, &column2).context(ArrowComputeSnafu)?;
-
-        let result = zip::zip(&boolean_array, &column1, &column2).context(ArrowComputeSnafu)?;
-        Helper::try_into_vector(&result).context(error::FromArrowArraySnafu)
-    }};
-}
-
-impl Function for GreatestFunction {
-    fn name(&self) -> &str {
-        NAME
-    }
-
-    fn return_type(&self, input_types: &[ConcreteDataType]) -> Result<ConcreteDataType> {
-        ensure!(
-            input_types.len() == 2,
-            InvalidFuncArgsSnafu {
-                err_msg: format!(
-                    "The length of the args is not correct, expect exactly two, have: {}",
-                    input_types.len()
-                )
-            }
-        );
-
-        match &input_types[0] {
-            ConcreteDataType::String(_) => Ok(ConcreteDataType::timestamp_millisecond_datatype()),
-            ConcreteDataType::Date(_) => Ok(ConcreteDataType::date_datatype()),
-            ConcreteDataType::Timestamp(ts_type) => Ok(ConcreteDataType::Timestamp(*ts_type)),
-            _ => UnsupportedInputDataTypeSnafu {
-                function: NAME,
-                datatypes: input_types,
-            }
-            .fail(),
-        }
-    }
-
-    fn signature(&self) -> Signature {
-        Signature::uniform(
-            2,
-            vec![
-                ConcreteDataType::string_datatype(),
-                ConcreteDataType::date_datatype(),
-                ConcreteDataType::timestamp_nanosecond_datatype(),
-                ConcreteDataType::timestamp_microsecond_datatype(),
-                ConcreteDataType::timestamp_millisecond_datatype(),
-                ConcreteDataType::timestamp_second_datatype(),
-            ],
-            Volatility::Immutable,
-        )
-    }
-
-    fn eval(&self, _func_ctx: &FunctionContext, columns: &[VectorRef]) -> Result<VectorRef> {
-        ensure!(
-            columns.len() == 2,
-            InvalidFuncArgsSnafu {
-                err_msg: format!(
-                    "The length of the args is not correct, expect exactly two, have: {}",
-                    columns.len()
-                ),
-            }
-        );
-        match columns[0].data_type() {
-            ConcreteDataType::String(_) => {
-                let column1 = cast(
-                    &columns[0].to_arrow_array(),
-                    &ArrowDataType::Timestamp(TimeUnit::Millisecond, None),
-                )
-                .context(ArrowComputeSnafu)?;
-                let column1 = column1.as_primitive::<TimestampMillisecondType>();
-                let column2 = cast(
-                    &columns[1].to_arrow_array(),
-                    &ArrowDataType::Timestamp(TimeUnit::Millisecond, None),
-                )
-                .context(ArrowComputeSnafu)?;
-                let column2 = column2.as_primitive::<TimestampMillisecondType>();
-                let boolean_array = gt(&column1, &column2).context(ArrowComputeSnafu)?;
-                let result =
-                    zip::zip(&boolean_array, &column1, &column2).context(ArrowComputeSnafu)?;
-                Ok(Helper::try_into_vector(&result).context(error::FromArrowArraySnafu)?)
-            }
-            ConcreteDataType::Date(_) => gt_time_types!(Date32Type, columns),
-            ConcreteDataType::Timestamp(ts_type) => match ts_type {
-                TimestampType::Second(_) => gt_time_types!(TimestampSecondType, columns),
-                TimestampType::Millisecond(_) => {
-                    gt_time_types!(TimestampMillisecondType, columns)
-                }
-                TimestampType::Microsecond(_) => {
-                    gt_time_types!(TimestampMicrosecondType, columns)
-                }
-                TimestampType::Nanosecond(_) => {
-                    gt_time_types!(TimestampNanosecondType, columns)
-                }
-            },
-            _ => UnsupportedInputDataTypeSnafu {
-                function: NAME,
-                datatypes: columns.iter().map(|c| c.data_type()).collect::<Vec<_>>(),
-            }
-            .fail(),
-        }
-    }
-}
-
-impl fmt::Display for GreatestFunction {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        write!(f, "GREATEST")
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use std::sync::Arc;
-
-    use common_time::timestamp::TimeUnit;
-    use common_time::{Date, Timestamp};
-    use datatypes::types::{
-        DateType, TimestampMicrosecondType, TimestampMillisecondType, TimestampNanosecondType,
-        TimestampSecondType,
-    };
-    use datatypes::value::Value;
-    use datatypes::vectors::{
-        DateVector, StringVector, TimestampMicrosecondVector, TimestampMillisecondVector,
-        TimestampNanosecondVector, TimestampSecondVector, Vector,
-    };
-    use paste::paste;
-
-    use super::*;
-    #[test]
-    fn test_greatest_takes_string_vector() {
-        let function = GreatestFunction;
-        assert_eq!(
-            function
-                .return_type(&[
-                    ConcreteDataType::string_datatype(),
-                    ConcreteDataType::string_datatype()
-                ])
-                .unwrap(),
-            ConcreteDataType::timestamp_millisecond_datatype()
-        );
-        let columns = vec![
-            Arc::new(StringVector::from(vec![
-                "1970-01-01".to_string(),
-                "2012-12-23".to_string(),
-            ])) as _,
-            Arc::new(StringVector::from(vec![
-                "2001-02-01".to_string(),
-                "1999-01-01".to_string(),
-            ])) as _,
-        ];
-
-        let result = function
-            .eval(&FunctionContext::default(), &columns)
-            .unwrap();
-        let result = result
-            .as_any()
-            .downcast_ref::<TimestampMillisecondVector>()
-            .unwrap();
-        assert_eq!(result.len(), 2);
-        assert_eq!(
-            result.get(0),
-            Value::Timestamp(Timestamp::from_str("2001-02-01 00:00:00", None).unwrap())
-        );
-        assert_eq!(
-            result.get(1),
-            Value::Timestamp(Timestamp::from_str("2012-12-23 00:00:00", None).unwrap())
-        );
-    }
-
-    #[test]
-    fn test_greatest_takes_date_vector() {
-        let function = GreatestFunction;
-        assert_eq!(
-            function
-                .return_type(&[
-                    ConcreteDataType::date_datatype(),
-                    ConcreteDataType::date_datatype()
-                ])
-                .unwrap(),
-            ConcreteDataType::Date(DateType)
-        );
-
-        let columns = vec![
-            Arc::new(DateVector::from_slice(vec![-1, 2])) as _,
-            Arc::new(DateVector::from_slice(vec![0, 1])) as _,
-        ];
-
-        let result = function
-            .eval(&FunctionContext::default(), &columns)
-            .unwrap();
-        let result = result.as_any().downcast_ref::<DateVector>().unwrap();
-        assert_eq!(result.len(), 2);
-        assert_eq!(
-            result.get(0),
-            Value::Date(Date::from_str_utc("1970-01-01").unwrap())
-        );
-        assert_eq!(
-            result.get(1),
-            Value::Date(Date::from_str_utc("1970-01-03").unwrap())
-        );
-    }
-
-    #[test]
-    fn test_greatest_takes_datetime_vector() {
-        let function = GreatestFunction;
-        assert_eq!(
-            function
-                .return_type(&[
-                    ConcreteDataType::timestamp_millisecond_datatype(),
-                    ConcreteDataType::timestamp_millisecond_datatype()
-                ])
-                .unwrap(),
-            ConcreteDataType::timestamp_millisecond_datatype()
-        );
-
-        let columns = vec![
-            Arc::new(TimestampMillisecondVector::from_slice(vec![-1, 2])) as _,
-            Arc::new(TimestampMillisecondVector::from_slice(vec![0, 1])) as _,
-        ];
-
-        let result = function
-            .eval(&FunctionContext::default(), &columns)
-            .unwrap();
-        let result = result
-            .as_any()
-            .downcast_ref::<TimestampMillisecondVector>()
-            .unwrap();
-        assert_eq!(result.len(), 2);
-        assert_eq!(
-            result.get(0),
-            Value::Timestamp(Timestamp::from_str("1970-01-01 00:00:00", None).unwrap())
-        );
-        assert_eq!(
-            result.get(1),
-            Value::Timestamp(Timestamp::from_str("1970-01-01 00:00:00.002", None).unwrap())
-        );
-    }
-
-    macro_rules! test_timestamp {
-        ($type: expr,$unit: ident) => {
-            paste! {
-                #[test]
-                fn [<test_greatest_takes_ $unit:lower _vector>]() {
-                    let function = GreatestFunction;
-                    assert_eq!(
-                        function.return_type(&[$type, $type]).unwrap(),
-                        ConcreteDataType::Timestamp(TimestampType::$unit([<Timestamp $unit Type>]))
-                    );
-
-                    let columns = vec![
-                        Arc::new([<Timestamp $unit Vector>]::from_slice(vec![-1, 2])) as _,
-                        Arc::new([<Timestamp $unit Vector>]::from_slice(vec![0, 1])) as _,
-                    ];
-
-                    let result = function.eval(&FunctionContext::default(), &columns).unwrap();
-                    let result = result.as_any().downcast_ref::<[<Timestamp $unit Vector>]>().unwrap();
-                    assert_eq!(result.len(), 2);
-                    assert_eq!(
-                        result.get(0),
-                        Value::Timestamp(Timestamp::new(0, TimeUnit::$unit))
-                    );
-                    assert_eq!(
-                        result.get(1),
-                        Value::Timestamp(Timestamp::new(2, TimeUnit::$unit))
-                    );
-                }
-            }
-        }
-    }
-
-    test_timestamp!(
-        ConcreteDataType::timestamp_nanosecond_datatype(),
-        Nanosecond
-    );
-    test_timestamp!(
-        ConcreteDataType::timestamp_microsecond_datatype(),
-        Microsecond
-    );
-    test_timestamp!(
-        ConcreteDataType::timestamp_millisecond_datatype(),
-        Millisecond
-    );
-    test_timestamp!(ConcreteDataType::timestamp_second_datatype(), Second);
-}

src/mito2/src/memtable/partition_tree.rs

@@ -302,7 +302,10 @@ impl PartitionTreeMemtable {
     fn update_stats(&self, metrics: &WriteMetrics) {
         // Only let the tracker tracks value bytes.
         self.alloc_tracker.on_allocation(metrics.value_bytes);
-        metrics.update_timestamp_range(&self.max_timestamp, &self.min_timestamp);
+        self.max_timestamp
+            .fetch_max(metrics.max_ts, Ordering::SeqCst);
+        self.min_timestamp
+            .fetch_min(metrics.min_ts, Ordering::SeqCst);
     }
 }
 

src/mito2/src/memtable/stats.rs

@@ -14,8 +14,6 @@
 
 //! Internal metrics of the memtable.
 
-use std::sync::atomic::{AtomicI64, Ordering};
-
 /// Metrics of writing memtables.
 pub(crate) struct WriteMetrics {
     /// Size allocated by keys.
@@ -28,51 +26,6 @@ pub(crate) struct WriteMetrics {
     pub(crate) max_ts: i64,
 }
 
-impl WriteMetrics {
-    /// Update the min/max timestamp range according to current write metric.
-    pub(crate) fn update_timestamp_range(&self, prev_max_ts: &AtomicI64, prev_min_ts: &AtomicI64) {
-        loop {
-            let current_min = prev_min_ts.load(Ordering::Relaxed);
-            if self.min_ts >= current_min {
-                break;
-            }
-
-            let Err(updated) = prev_min_ts.compare_exchange(
-                current_min,
-                self.min_ts,
-                Ordering::Relaxed,
-                Ordering::Relaxed,
-            ) else {
-                break;
-            };
-
-            if updated == self.min_ts {
-                break;
-            }
-        }
-
-        loop {
-            let current_max = prev_max_ts.load(Ordering::Relaxed);
-            if self.max_ts <= current_max {
-                break;
-            }
-
-            let Err(updated) = prev_max_ts.compare_exchange(
-                current_max,
-                self.max_ts,
-                Ordering::Relaxed,
-                Ordering::Relaxed,
-            ) else {
-                break;
-            };
-
-            if updated == self.max_ts {
-                break;
-            }
-        }
-    }
-}
-
 impl Default for WriteMetrics {
     fn default() -> Self {
         Self {

src/mito2/src/memtable/time_series.rs

@@ -147,7 +147,8 @@ impl TimeSeriesMemtable {
     fn update_stats(&self, stats: WriteMetrics) {
         self.alloc_tracker
             .on_allocation(stats.key_bytes + stats.value_bytes);
-        stats.update_timestamp_range(&self.max_timestamp, &self.min_timestamp);
+        self.max_timestamp.fetch_max(stats.max_ts, Ordering::SeqCst);
+        self.min_timestamp.fetch_min(stats.min_ts, Ordering::SeqCst);
     }
 
     fn write_key_value(&self, kv: KeyValue, stats: &mut WriteMetrics) -> Result<()> {

src/mito2/src/read/scan_region.rs

@@ -322,13 +322,10 @@ impl ScanRegion {
         let memtables: Vec<_> = memtables
             .into_iter()
             .filter(|mem| {
-                if mem.is_empty() {
+                // check if memtable is empty by reading stats.
+                let Some((start, end)) = mem.stats().time_range() else {
                     return false;
-                }
+                };
-                let stats = mem.stats();
-                // Safety: the memtable is not empty.
-                let (start, end) = stats.time_range().unwrap();
-
                 // The time range of the memtable is inclusive.
                 let memtable_range = TimestampRange::new_inclusive(Some(start), Some(end));
                 memtable_range.intersects(&time_range)

src/mito2/src/sst/parquet/format.rs

@@ -134,6 +134,7 @@ impl WriteFormat {
 
 /// Helper for reading the SST format.
 pub struct ReadFormat {
+    /// The metadata stored in the SST.
     metadata: RegionMetadataRef,
     /// SST file schema.
     arrow_schema: SchemaRef,
@@ -305,17 +306,23 @@ impl ReadFormat {
         &self,
         row_groups: &[impl Borrow<RowGroupMetaData>],
         column_id: ColumnId,
-    ) -> Option<ArrayRef> {
+    ) -> StatValues {
-        let column = self.metadata.column_by_id(column_id)?;
+        let Some(column) = self.metadata.column_by_id(column_id) else {
+            // No such column in the SST.
+            return StatValues::NoColumn;
+        };
         match column.semantic_type {
             SemanticType::Tag => self.tag_values(row_groups, column, true),
             SemanticType::Field => {
-                let index = self.field_id_to_index.get(&column_id)?;
+                // Safety: `field_id_to_index` is initialized by the semantic type.
-                Self::column_values(row_groups, column, *index, true)
+                let index = self.field_id_to_index.get(&column_id).unwrap();
+                let stats = Self::column_values(row_groups, column, *index, true);
+                StatValues::from_stats_opt(stats)
             }
             SemanticType::Timestamp => {
                 let index = self.time_index_position();
-                Self::column_values(row_groups, column, index, true)
+                let stats = Self::column_values(row_groups, column, index, true);
+                StatValues::from_stats_opt(stats)
             }
         }
     }
@@ -325,17 +332,23 @@ impl ReadFormat {
         &self,
         row_groups: &[impl Borrow<RowGroupMetaData>],
         column_id: ColumnId,
-    ) -> Option<ArrayRef> {
+    ) -> StatValues {
-        let column = self.metadata.column_by_id(column_id)?;
+        let Some(column) = self.metadata.column_by_id(column_id) else {
+            // No such column in the SST.
+            return StatValues::NoColumn;
+        };
         match column.semantic_type {
             SemanticType::Tag => self.tag_values(row_groups, column, false),
             SemanticType::Field => {
-                let index = self.field_id_to_index.get(&column_id)?;
+                // Safety: `field_id_to_index` is initialized by the semantic type.
-                Self::column_values(row_groups, column, *index, false)
+                let index = self.field_id_to_index.get(&column_id).unwrap();
+                let stats = Self::column_values(row_groups, column, *index, false);
+                StatValues::from_stats_opt(stats)
             }
             SemanticType::Timestamp => {
                 let index = self.time_index_position();
-                Self::column_values(row_groups, column, index, false)
+                let stats = Self::column_values(row_groups, column, index, false);
+                StatValues::from_stats_opt(stats)
             }
         }
     }
@@ -345,17 +358,23 @@ impl ReadFormat {
         &self,
         row_groups: &[impl Borrow<RowGroupMetaData>],
         column_id: ColumnId,
-    ) -> Option<ArrayRef> {
+    ) -> StatValues {
-        let column = self.metadata.column_by_id(column_id)?;
+        let Some(column) = self.metadata.column_by_id(column_id) else {
+            // No such column in the SST.
+            return StatValues::NoColumn;
+        };
         match column.semantic_type {
-            SemanticType::Tag => None,
+            SemanticType::Tag => StatValues::NoStats,
             SemanticType::Field => {
-                let index = self.field_id_to_index.get(&column_id)?;
+                // Safety: `field_id_to_index` is initialized by the semantic type.
-                Self::column_null_counts(row_groups, *index)
+                let index = self.field_id_to_index.get(&column_id).unwrap();
+                let stats = Self::column_null_counts(row_groups, *index);
+                StatValues::from_stats_opt(stats)
             }
             SemanticType::Timestamp => {
                 let index = self.time_index_position();
-                Self::column_null_counts(row_groups, index)
+                let stats = Self::column_null_counts(row_groups, index);
+                StatValues::from_stats_opt(stats)
             }
         }
     }
@@ -390,8 +409,7 @@ impl ReadFormat {
         row_groups: &[impl Borrow<RowGroupMetaData>],
         column: &ColumnMetadata,
         is_min: bool,
-    ) -> Option<ArrayRef> {
+    ) -> StatValues {
-        let primary_key_encoding = self.metadata.primary_key_encoding;
         let is_first_tag = self
             .metadata
             .primary_key
@@ -400,9 +418,28 @@ impl ReadFormat {
             .unwrap_or(false);
         if !is_first_tag {
             // Only the min-max of the first tag is available in the primary key.
-            return None;
+            return StatValues::NoStats;
         }
 
+        StatValues::from_stats_opt(self.first_tag_values(row_groups, column, is_min))
+    }
+
+    /// Returns min/max values of the first tag.
+    /// Returns None if the tag does not have statistics.
+    fn first_tag_values(
+        &self,
+        row_groups: &[impl Borrow<RowGroupMetaData>],
+        column: &ColumnMetadata,
+        is_min: bool,
+    ) -> Option<ArrayRef> {
+        debug_assert!(self
+            .metadata
+            .primary_key
+            .first()
+            .map(|id| *id == column.column_id)
+            .unwrap_or(false));
+
+        let primary_key_encoding = self.metadata.primary_key_encoding;
         let converter = build_primary_key_codec_with_fields(
             primary_key_encoding,
             [(
@@ -452,6 +489,7 @@ impl ReadFormat {
     }
 
     /// Returns min/max values of specific non-tag columns.
+    /// Returns None if the column does not have statistics.
     fn column_values(
         row_groups: &[impl Borrow<RowGroupMetaData>],
         column: &ColumnMetadata,
@@ -544,6 +582,29 @@ impl ReadFormat {
     }
 }
 
+/// Values of column statistics of the SST.
+///
+/// It also distinguishes the case that a column is not found and
+/// the column exists but has no statistics.
+pub enum StatValues {
+    /// Values of each row group.
+    Values(ArrayRef),
+    /// No such column.
+    NoColumn,
+    /// Column exists but has no statistics.
+    NoStats,
+}
+
+impl StatValues {
+    /// Creates a new `StatValues` instance from optional statistics.
+    pub fn from_stats_opt(stats: Option<ArrayRef>) -> Self {
+        match stats {
+            Some(stats) => StatValues::Values(stats),
+            None => StatValues::NoStats,
+        }
+    }
+}
+
 #[cfg(test)]
 impl ReadFormat {
     /// Creates a helper with existing `metadata` and all columns.

src/mito2/src/sst/parquet/stats.rs

@@ -25,7 +25,7 @@ use parquet::file::metadata::RowGroupMetaData;
 use store_api::metadata::RegionMetadataRef;
 use store_api::storage::ColumnId;
 
-use crate::sst::parquet::format::ReadFormat;
+use crate::sst::parquet::format::{ReadFormat, StatValues};
 
 /// Statistics for pruning row groups.
 pub(crate) struct RowGroupPruningStats<'a, T> {
@@ -100,16 +100,18 @@ impl<T: Borrow<RowGroupMetaData>> PruningStatistics for RowGroupPruningStats<'_,
     fn min_values(&self, column: &Column) -> Option<ArrayRef> {
         let column_id = self.column_id_to_prune(&column.name)?;
         match self.read_format.min_values(self.row_groups, column_id) {
-            Some(values) => Some(values),
+            StatValues::Values(values) => Some(values),
-            None => self.compat_default_value(&column.name),
+            StatValues::NoColumn => self.compat_default_value(&column.name),
+            StatValues::NoStats => None,
         }
     }
 
     fn max_values(&self, column: &Column) -> Option<ArrayRef> {
         let column_id = self.column_id_to_prune(&column.name)?;
         match self.read_format.max_values(self.row_groups, column_id) {
-            Some(values) => Some(values),
+            StatValues::Values(values) => Some(values),
-            None => self.compat_default_value(&column.name),
+            StatValues::NoColumn => self.compat_default_value(&column.name),
+            StatValues::NoStats => None,
         }
     }
 
@@ -118,10 +120,12 @@ impl<T: Borrow<RowGroupMetaData>> PruningStatistics for RowGroupPruningStats<'_,
     }
 
     fn null_counts(&self, column: &Column) -> Option<ArrayRef> {
-        let Some(column_id) = self.column_id_to_prune(&column.name) else {
+        let column_id = self.column_id_to_prune(&column.name)?;
-            return self.compat_null_count(&column.name);
+        match self.read_format.null_counts(self.row_groups, column_id) {
-        };
+            StatValues::Values(values) => Some(values),
-        self.read_format.null_counts(self.row_groups, column_id)
+            StatValues::NoColumn => self.compat_null_count(&column.name),
+            StatValues::NoStats => None,
+        }
     }
 
     fn row_counts(&self, _column: &Column) -> Option<ArrayRef> {

tests/cases/standalone/common/flow/flow_step_aggr.result

@@ -50,6 +50,7 @@ ADMIN FLUSH_FLOW('calc_access_log_10s');
 +-----------------------------------------+
 
 -- query should return 3 rows
+-- SQLNESS SORT_RESULT 3 1
 SELECT "url", time_window FROM access_log_10s
 ORDER BY
     time_window;
@@ -63,6 +64,7 @@ ORDER BY
 +------------+---------------------+
 
 -- use hll_count to query the approximate data in access_log_10s
+-- SQLNESS SORT_RESULT 3 1
 SELECT "url", time_window, hll_count(state) FROM access_log_10s
 ORDER BY
     time_window;
@@ -76,6 +78,7 @@ ORDER BY
 +------------+---------------------+---------------------------------+
 
 -- further, we can aggregate 10 seconds of data to every minute, by using hll_merge to merge 10 seconds of hyperloglog state
+-- SQLNESS SORT_RESULT 3 1
 SELECT
     "url",
     date_bin('1 minute'::INTERVAL, time_window) AS time_window_1m,
@@ -91,8 +94,8 @@ ORDER BY
 +------------+---------------------+------------+
 | url        | time_window_1m      | uv_per_min |
 +------------+---------------------+------------+
-| /not_found | 2025-03-04T00:00:00 | 1          |
 | /dashboard | 2025-03-04T00:00:00 | 3          |
+| /not_found | 2025-03-04T00:00:00 | 1          |
 +------------+---------------------+------------+
 
 DROP FLOW calc_access_log_10s;

tests/cases/standalone/common/flow/flow_step_aggr.sql

@@ -36,16 +36,19 @@ INSERT INTO access_log VALUES
 ADMIN FLUSH_FLOW('calc_access_log_10s');
 
 -- query should return 3 rows
+-- SQLNESS SORT_RESULT 3 1
 SELECT "url", time_window FROM access_log_10s
 ORDER BY
     time_window;
 
 -- use hll_count to query the approximate data in access_log_10s
+-- SQLNESS SORT_RESULT 3 1
 SELECT "url", time_window, hll_count(state) FROM access_log_10s
 ORDER BY
     time_window;
 
 -- further, we can aggregate 10 seconds of data to every minute, by using hll_merge to merge 10 seconds of hyperloglog state
+-- SQLNESS SORT_RESULT 3 1
 SELECT
     "url",
     date_bin('1 minute'::INTERVAL, time_window) AS time_window_1m,

tests/cases/standalone/common/function/time.result

@@ -9,7 +9,7 @@ select GREATEST('1999-01-30', '2023-03-01');
 +-------------------------------------------------+
 | greatest(Utf8("1999-01-30"),Utf8("2023-03-01")) |
 +-------------------------------------------------+
-| 2023-03-01T00:00:00                             |
+| 2023-03-01                                      |
 +-------------------------------------------------+
 
 select GREATEST('2000-02-11'::Date, '2020-12-30'::Date);

tests/cases/standalone/common/select/prune_pk.result

@@ -0,0 +1,158 @@
+CREATE TABLE IF NOT EXISTS `test_multi_pk_filter` ( `namespace` STRING NULL, `env` STRING NULL DEFAULT 'NULL', `flag` INT NULL, `total` BIGINT NULL, `greptime_timestamp` TIMESTAMP(9) NOT NULL, TIME INDEX (`greptime_timestamp`), PRIMARY KEY (`namespace`, `env`, `flag`) ) ENGINE=mito;
+
+Affected Rows: 0
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 1, 5289, '2023-05-15 10:00:00');
+
+Affected Rows: 1
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 0, 421, '2023-05-15 10:05:00');
+
+Affected Rows: 1
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'dev', 1, 356, '2023-05-15 10:10:00');
+
+Affected Rows: 1
+
+ADMIN FLUSH_TABLE('test_multi_pk_filter');
+
++-------------------------------------------+
+| ADMIN FLUSH_TABLE('test_multi_pk_filter') |
++-------------------------------------------+
+| 0                                         |
++-------------------------------------------+
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'dev', 1, 412, '2023-05-15 10:15:00');
+
+Affected Rows: 1
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'dev', 1, 298, '2023-05-15 10:20:00');
+
+Affected Rows: 1
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 1, 5289, '2023-05-15 10:25:00');
+
+Affected Rows: 1
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 1, 5874, '2023-05-15 10:30:00');
+
+Affected Rows: 1
+
+ADMIN FLUSH_TABLE('test_multi_pk_filter');
+
++-------------------------------------------+
+| ADMIN FLUSH_TABLE('test_multi_pk_filter') |
++-------------------------------------------+
+| 0                                         |
++-------------------------------------------+
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 1, 6132, '2023-05-15 10:35:00');
+
+Affected Rows: 1
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'testing', 1, 1287, '2023-05-15 10:40:00');
+
+Affected Rows: 1
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'testing', 1, 1432, '2023-05-15 10:45:00');
+
+Affected Rows: 1
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'testing', 1, 1056, '2023-05-15 10:50:00');
+
+Affected Rows: 1
+
+SELECT greptime_timestamp, namespace, env, total FROM test_multi_pk_filter WHERE
+    greptime_timestamp BETWEEN '2023-05-15 10:00:00' AND '2023-05-15 11:00:00' AND flag = 1 AND namespace = 'thermostat_v2'
+    ORDER BY greptime_timestamp;
+
++---------------------+---------------+------------+-------+
+| greptime_timestamp  | namespace     | env        | total |
++---------------------+---------------+------------+-------+
+| 2023-05-15T10:00:00 | thermostat_v2 | production | 5289  |
+| 2023-05-15T10:10:00 | thermostat_v2 | dev        | 356   |
+| 2023-05-15T10:15:00 | thermostat_v2 | dev        | 412   |
+| 2023-05-15T10:20:00 | thermostat_v2 | dev        | 298   |
+| 2023-05-15T10:25:00 | thermostat_v2 | production | 5289  |
+| 2023-05-15T10:30:00 | thermostat_v2 | production | 5874  |
+| 2023-05-15T10:35:00 | thermostat_v2 | production | 6132  |
+| 2023-05-15T10:40:00 | thermostat_v2 | testing    | 1287  |
+| 2023-05-15T10:45:00 | thermostat_v2 | testing    | 1432  |
+| 2023-05-15T10:50:00 | thermostat_v2 | testing    | 1056  |
++---------------------+---------------+------------+-------+
+
+SELECT greptime_timestamp, namespace, env, total FROM test_multi_pk_filter WHERE
+    greptime_timestamp BETWEEN '2023-05-15 10:00:00' AND '2023-05-15 11:00:00' AND flag = 1 AND namespace = 'thermostat_v2' AND env='dev'
+    ORDER BY greptime_timestamp;
+
++---------------------+---------------+-----+-------+
+| greptime_timestamp  | namespace     | env | total |
++---------------------+---------------+-----+-------+
+| 2023-05-15T10:10:00 | thermostat_v2 | dev | 356   |
+| 2023-05-15T10:15:00 | thermostat_v2 | dev | 412   |
+| 2023-05-15T10:20:00 | thermostat_v2 | dev | 298   |
++---------------------+---------------+-----+-------+
+
+DROP TABLE test_multi_pk_filter;
+
+Affected Rows: 0
+
+CREATE TABLE IF NOT EXISTS `test_multi_pk_null` ( `namespace` STRING NULL, `env` STRING NULL DEFAULT 'NULL', `total` BIGINT NULL, `greptime_timestamp` TIMESTAMP(9) NOT NULL, TIME INDEX (`greptime_timestamp`), PRIMARY KEY (`namespace`, `env`) ) ENGINE=mito;
+
+Affected Rows: 0
+
+INSERT INTO test_multi_pk_null
+    (namespace, env, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 5289, '2023-05-15 10:00:00');
+
+Affected Rows: 1
+
+INSERT INTO test_multi_pk_null
+    (namespace, env, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 421, '2023-05-15 10:05:00');
+
+Affected Rows: 1
+
+ADMIN FLUSH_TABLE('test_multi_pk_null');
+
++-----------------------------------------+
+| ADMIN FLUSH_TABLE('test_multi_pk_null') |
++-----------------------------------------+
+| 0                                       |
++-----------------------------------------+
+
+SELECT * FROM test_multi_pk_null WHERE env IS NOT NULL;
+
++---------------+------------+-------+---------------------+
+| namespace     | env        | total | greptime_timestamp  |
++---------------+------------+-------+---------------------+
+| thermostat_v2 | production | 5289  | 2023-05-15T10:00:00 |
+| thermostat_v2 | production | 421   | 2023-05-15T10:05:00 |
++---------------+------------+-------+---------------------+
+
+DROP TABLE test_multi_pk_null;
+
+Affected Rows: 0
+

tests/cases/standalone/common/select/prune_pk.sql

@@ -0,0 +1,66 @@
+CREATE TABLE IF NOT EXISTS `test_multi_pk_filter` ( `namespace` STRING NULL, `env` STRING NULL DEFAULT 'NULL', `flag` INT NULL, `total` BIGINT NULL, `greptime_timestamp` TIMESTAMP(9) NOT NULL, TIME INDEX (`greptime_timestamp`), PRIMARY KEY (`namespace`, `env`, `flag`) ) ENGINE=mito;
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 1, 5289, '2023-05-15 10:00:00');
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 0, 421, '2023-05-15 10:05:00');
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'dev', 1, 356, '2023-05-15 10:10:00');
+
+ADMIN FLUSH_TABLE('test_multi_pk_filter');
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'dev', 1, 412, '2023-05-15 10:15:00');
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'dev', 1, 298, '2023-05-15 10:20:00');
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 1, 5289, '2023-05-15 10:25:00');
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 1, 5874, '2023-05-15 10:30:00');
+
+ADMIN FLUSH_TABLE('test_multi_pk_filter');
+
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 1, 6132, '2023-05-15 10:35:00');
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'testing', 1, 1287, '2023-05-15 10:40:00');
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'testing', 1, 1432, '2023-05-15 10:45:00');
+INSERT INTO test_multi_pk_filter
+    (namespace, env, flag, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'testing', 1, 1056, '2023-05-15 10:50:00');
+
+SELECT greptime_timestamp, namespace, env, total FROM test_multi_pk_filter WHERE
+    greptime_timestamp BETWEEN '2023-05-15 10:00:00' AND '2023-05-15 11:00:00' AND flag = 1 AND namespace = 'thermostat_v2'
+    ORDER BY greptime_timestamp;
+
+SELECT greptime_timestamp, namespace, env, total FROM test_multi_pk_filter WHERE
+    greptime_timestamp BETWEEN '2023-05-15 10:00:00' AND '2023-05-15 11:00:00' AND flag = 1 AND namespace = 'thermostat_v2' AND env='dev'
+    ORDER BY greptime_timestamp;
+
+DROP TABLE test_multi_pk_filter;
+
+CREATE TABLE IF NOT EXISTS `test_multi_pk_null` ( `namespace` STRING NULL, `env` STRING NULL DEFAULT 'NULL', `total` BIGINT NULL, `greptime_timestamp` TIMESTAMP(9) NOT NULL, TIME INDEX (`greptime_timestamp`), PRIMARY KEY (`namespace`, `env`) ) ENGINE=mito;
+
+INSERT INTO test_multi_pk_null
+    (namespace, env, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 5289, '2023-05-15 10:00:00');
+INSERT INTO test_multi_pk_null
+    (namespace, env, total, greptime_timestamp)
+    VALUES ('thermostat_v2', 'production', 421, '2023-05-15 10:05:00');
+
+ADMIN FLUSH_TABLE('test_multi_pk_null');
+
+SELECT * FROM test_multi_pk_null WHERE env IS NOT NULL;
+
+DROP TABLE test_multi_pk_null;