perf: set simple filter on primary key columns to exact filter (Grept…

…imeTeam#4564) * perf: set simple filter on primary key columns to exact filter Signed-off-by: Ruihang Xia <[email protected]> * add sqlness test Signed-off-by: Ruihang Xia <[email protected]> * fix typo Signed-off-by: Ruihang Xia <[email protected]> * fix sqlness Signed-off-by: Ruihang Xia <[email protected]> --------- Signed-off-by: Ruihang Xia <[email protected]>
lyang24 · Aug 19, 2024 · 8de11a0 · 8de11a0
1 parent 975b8c6
commit 8de11a0
Show file tree

Hide file tree

Showing 3 changed files with 47 additions and 1 deletion.
diff --git a/src/query/src/dummy_catalog.rs b/src/query/src/dummy_catalog.rs
@@ -17,7 +17,9 @@
 use std::any::Any;
 use std::sync::{Arc, Mutex};
 
+use api::v1::SemanticType;
 use async_trait::async_trait;
+use common_recordbatch::filter::SimpleFilterEvaluator;
 use common_recordbatch::OrderOption;
 use datafusion::catalog::schema::SchemaProvider;
 use datafusion::catalog::{CatalogProvider, CatalogProviderList};
@@ -177,7 +179,27 @@ impl TableProvider for DummyTableProvider {
         &self,
         filters: &[&Expr],
     ) -> datafusion::error::Result<Vec<TableProviderFilterPushDown>> {
-        Ok(vec![TableProviderFilterPushDown::Inexact; filters.len()])
+        let supported = filters
+            .iter()
+            .map(|e| {
+                // Simple filter on primary key columns are precisely evaluated.
+                if let Some(simple_filter) = SimpleFilterEvaluator::try_new(e) {
+                    if self
+                        .metadata
+                        .column_by_name(simple_filter.column_name())
+                        .and_then(|c| (c.semantic_type == SemanticType::Tag).then_some(()))
+                        .is_some()
+                    {
+                        TableProviderFilterPushDown::Exact
+                    } else {
+                        TableProviderFilterPushDown::Inexact
+                    }
+                } else {
+                    TableProviderFilterPushDown::Inexact
+                }
+            })
+            .collect();
+        Ok(supported)
     }
 }
 

diff --git a/tests/cases/standalone/common/select/prune.result b/tests/cases/standalone/common/select/prune.result
@@ -77,6 +77,23 @@ select * from demo where collector='disk' order by ts;
 | 1970-01-01T00:00:00.002 | 3.0   | test2 | idc1 | disk      |
 +-------------------------+-------+-------+------+-----------+
 
+-- SQLNESS REPLACE (-+) -
+-- SQLNESS REPLACE (\s\s+) _
+-- SQLNESS REPLACE (peers.*) REDACTED
+-- SQLNESS REPLACE (metrics.*) REDACTED
+-- SQLNESS REPLACE region=\d+\(\d+,\s+\d+\) region=REDACTED
+explain analyze select * from demo where idc='idc1';
+
++-+-+-+
+| stage | node | plan_|
++-+-+-+
+| 0_| 0_|_MergeScanExec: REDACTED
+|_|_|_|
+| 1_| 0_|_SeqScan: region=REDACTED, partition_count=1 (1 memtable ranges, 0 file ranges) REDACTED
+|_|_|_|
+|_|_| Total rows: 2_|
++-+-+-+
+
 drop table demo;
 
 Affected Rows: 0

diff --git a/tests/cases/standalone/common/select/prune.sql b/tests/cases/standalone/common/select/prune.sql
@@ -20,4 +20,11 @@ select * from demo where idc='idc1' order by ts;
 
 select * from demo where collector='disk' order by ts;
 
+-- SQLNESS REPLACE (-+) -
+-- SQLNESS REPLACE (\s\s+) _
+-- SQLNESS REPLACE (peers.*) REDACTED
+-- SQLNESS REPLACE (metrics.*) REDACTED
+-- SQLNESS REPLACE region=\d+\(\d+,\s+\d+\) region=REDACTED
+explain analyze select * from demo where idc='idc1';
+
 drop table demo;