feat(query): add settings enable_selector_executor (#18434)

Author: sundy-li

src/query/expression/src/filter/filter_executor.rs

@@ -19,6 +19,7 @@ use databend_common_exception::Result;
 
 use crate::filter::SelectExpr;
 use crate::filter::Selector;
+use crate::types::BooleanType;
 use crate::ColumnSet;
 use crate::DataBlock;
 use crate::Evaluator;
@@ -40,6 +41,8 @@ pub struct FilterExecutor {
     selection_range: Vec<Range<u32>>,
     fn_registry: &'static FunctionRegistry,
     keep_order: bool,
+
+    filter: Expr,
 }
 
 impl FilterExecutor {
@@ -59,6 +62,7 @@ impl FilterExecutor {
         } else {
             vec![]
         };
+
         Self {
             select_expr,
             func_ctx,
@@ -70,14 +74,24 @@ impl FilterExecutor {
             selection_range: vec![],
             fn_registry,
             keep_order,
+            filter: expr,
         }
     }
 
     // Filter a DataBlock, return the filtered DataBlock.
     pub fn filter(&mut self, data_block: DataBlock) -> Result<DataBlock> {
-        let origin_count = data_block.num_rows();
-        let result_count = self.select(&data_block)?;
-        self.take(data_block, origin_count, result_count)
+        if self.func_ctx.enable_selector_executor {
+            let origin_count = data_block.num_rows();
+            let result_count = self.select(&data_block)?;
+            self.take(data_block, origin_count, result_count)
+        } else {
+            let evaluator = Evaluator::new(&data_block, &self.func_ctx, self.fn_registry);
+            let filter = evaluator
+                .run(&self.filter)?
+                .try_downcast::<BooleanType>()
+                .unwrap();
+            data_block.filter_boolean_value(&filter)
+        }
     }
 
     // Store the filtered indices of data_block in `true_selection` and return the number of filtered indices.

src/query/expression/src/function.rs

@@ -161,6 +161,7 @@ pub struct FunctionContext {
     pub now: Zoned,
     pub rounding_mode: bool,
     pub disable_variant_check: bool,
+    pub enable_selector_executor: bool,
 
     pub openai_api_chat_base_url: String,
     pub openai_api_embedding_base_url: String,
@@ -184,6 +185,7 @@ impl Default for FunctionContext {
             now: Default::default(),
             rounding_mode: false,
             disable_variant_check: false,
+            enable_selector_executor: true,
             openai_api_chat_base_url: "".to_string(),
             openai_api_embedding_base_url: "".to_string(),
             openai_api_key: "".to_string(),

src/query/service/src/sessions/query_ctx.rs

@@ -1014,6 +1014,7 @@ impl TableContext for QueryContext {
             tz,
             rounding_mode,
             disable_variant_check,
+            enable_selector_executor: settings.get_enable_selector_executor()?,
 
             openai_api_key: query_config.openai_api_key.clone(),
             openai_api_version: query_config.openai_api_version.clone(),

src/query/settings/src/settings_default.rs

@@ -1192,6 +1192,13 @@ impl DefaultSettings {
                     scope: SettingScope::Both,
                     range: Some(SettingRange::Numeric(0..=1)),
                 }),
+                ("enable_selector_executor", DefaultSettingValue {
+                    value: UserSettingValue::UInt64(1),
+                    desc: "Enables selector executor for filter expression",
+                    mode: SettingMode::Both,
+                    scope: SettingScope::Both,
+                    range: Some(SettingRange::Numeric(0..=1)),
+                }),
                 ("dynamic_sample_time_budget_ms", DefaultSettingValue {
                     value: UserSettingValue::UInt64(0),
                     desc: "Time budget for dynamic sample in milliseconds",

src/query/settings/src/settings_getter_setter.rs

@@ -865,6 +865,10 @@ impl Settings {
         Ok(self.try_get_u64("random_function_seed")? == 1)
     }
 
+    pub fn get_enable_selector_executor(&self) -> Result<bool> {
+        Ok(self.try_get_u64("enable_selector_executor")? == 1)
+    }
+
     pub fn get_dynamic_sample_time_budget_ms(&self) -> Result<u64> {
         self.try_get_u64("dynamic_sample_time_budget_ms")
     }

tests/sqllogictests/suites/query/filter.test

@@ -3,7 +3,7 @@ statement ok
 drop table if exists t;
 
 statement ok
-create table t(a int null, b varchar null);
+create or replace table t(a int null, b varchar null);
 
 statement ok
 insert into t(a, b) values (3, NULL), (3,'hello'), (3,'hello'), (3,'hello'), (3,'hello'), (3,'hello'), (3,'hello'), (3,'hello'), (1,'4'), (1,'4');
@@ -14,28 +14,10 @@ select a, b from t where a = 1 and cast(b as int) >= 2;
 1 4
 1 4
 
-# OR filter short circuit
-statement ok
-drop table if exists t;
-
-statement ok
-create table t(a int);
-
-statement ok
-insert into t values(0), (1), (2), (3), (NULL)
-
-query I
-select a from t where a = 0 or 3 / a > 2 order by a
-----
-0
-1
-
-statement ok
-drop table if exists t;
 
 # AND filter short circuit
 statement ok
-create table t(a varchar);
+create or replace table t(a varchar);
 
 statement ok
 insert into t values('null'), ('202405');
@@ -45,15 +27,9 @@ SELECT count(1) FROM t WHERE a <> 'null' AND a IS NOT NULL AND to_date(a || '01'
 ----
 1
 
-statement ok
-drop table if exists t;
 
-# Boolean comparison
 statement ok
-drop table if exists t;
-
-statement ok
-create table t(a boolean, b boolean);
+create or replace table t(a boolean, b boolean);
 
 statement ok
 insert into t values(true, true), (true, false), (false, false);
@@ -101,14 +77,7 @@ select * from t where not(not(not(a)))
 0 0
 
 statement ok
-drop table if exists t;
-
-# Like Test
-statement ok
-drop table if exists t;
-
-statement ok
-create table t(a int, b varchar);
+create or replace table t(a int, b varchar);
 
 statement ok
 insert into t values(1, 'databend'), (2, ' databend query'), (3, 'query databend'), (4, 'query databend, query'), (5, 'databend cloud data warehouse'), (6, 'databend%cloud data warehouse');
@@ -216,8 +185,6 @@ select * from t where _origin_version is null;
 1
 2
 
-statement ok
-drop table if exists t;
 
 #issue 17388
 
@@ -232,5 +199,9 @@ SELECT * FROM t2 WHERE payload:key2 = 'null';
 ----
 2 {"key":"val2","key2":"null"}
 
+
+statement ok
+drop table if exists t;
+
 statement ok
 drop table if exists t2;

tests/sqllogictests/suites/query/filter_disable_selector.test

@@ -0,0 +1,23 @@
+statement ok
+set enable_selector_executor = 0;
+
+include ./filter.test
+
+statement ok
+unset enable_selector_executor;
+
+
+statement ok
+create or replace table t(a int);
+
+statement ok
+insert into t values(0), (1), (2), (3), (NULL)
+
+query I
+select a from t where a = 0 or 3 / a > 2 order by a
+----
+0
+1
+
+statement ok
+drop table if exists t;