chore(query): turn on new agg hashtable (databendlabs#15155)

yufan022 · Jun 18, 2024 · e63acd0 · e63acd0
1 parent f2e7502
commit e63acd0
Show file tree

Hide file tree

Showing 5 changed files with 72 additions and 12 deletions.
diff --git a/src/query/expression/src/aggregate/aggregate_hashtable.rs b/src/query/expression/src/aggregate/aggregate_hashtable.rs
@@ -220,9 +220,11 @@ impl AggregateHashTable {
             }
         }
 
-        if self.config.partial_agg && self.capacity >= self.config.max_partial_capacity {
+        if self.config.partial_agg {
             // check size
-            if self.count + BATCH_ADD_SIZE > self.resize_threshold() {
+            if self.count + BATCH_ADD_SIZE > self.resize_threshold()
+                && self.capacity >= self.config.max_partial_capacity
+            {
                 self.clear_ht();
                 self.reset_count();
             }

diff --git a/src/query/settings/src/settings_default.rs b/src/query/settings/src/settings_default.rs
@@ -596,10 +596,10 @@ impl DefaultSettings {
                     range: Some(SettingRange::Numeric(0..=1)),
                 }),
                 ("enable_experimental_aggregate_hashtable", DefaultSettingValue {
-                        value: UserSettingValue::UInt64(0),
-                        desc: "Enables experimental aggregate hashtable",
-                        mode: SettingMode::Both,
-                        range: Some(SettingRange::Numeric(0..=1)),
+                    value: UserSettingValue::UInt64(1),
+                    desc: "Enables experimental aggregate hashtable",
+                    mode: SettingMode::Both,
+                    range: Some(SettingRange::Numeric(0..=1)),
                 }),
                 ("numeric_cast_option", DefaultSettingValue {
                     value: UserSettingValue::String("rounding".to_string()),

diff --git a/tests/sqllogictests/suites/mode/cluster/exchange.test b/tests/sqllogictests/suites/mode/cluster/exchange.test
@@ -167,8 +167,8 @@ Exchange
     │       ├── aggregate functions: [sum(number)]
     │       ├── estimated rows: 1.00
     │       └── Exchange
-    │           ├── output columns: [sum(number) (#2), #_group_by_key]
-    │           ├── exchange type: Hash(_group_by_key)
+    │           ├── output columns: [sum(number) (#2), numbers.number (#0)]
+    │           ├── exchange type: Hash(0)
     │           └── AggregatePartial
     │               ├── group by: [number]
     │               ├── aggregate functions: [sum(number)]
@@ -198,7 +198,7 @@ explain fragments select * from (select sum(number) as number from numbers(1) gr
 Fragment 0:
   DataExchange: Shuffle
     ExchangeSink
-    ├── output columns: [sum(number) (#2), #_group_by_key]
+    ├── output columns: [sum(number) (#2), numbers.number (#0)]
     ├── destination fragment: [1]
     └── AggregatePartial
         ├── group by: [number]
@@ -226,7 +226,7 @@ Fragment 1:
         ├── aggregate functions: [sum(number)]
         ├── estimated rows: 1.00
         └── ExchangeSource
-            ├── output columns: [sum(number) (#2), #_group_by_key]
+            ├── output columns: [sum(number) (#2), numbers.number (#0)]
             └── source fragment: [0]
 (empty)
 (empty)

diff --git a/tests/sqllogictests/suites/mode/cluster/explain_v2.test b/tests/sqllogictests/suites/mode/cluster/explain_v2.test
@@ -376,3 +376,61 @@ drop table t1;
 
 statement ok
 drop table t2;
+
+statement ok
+create table t1(a int, b int);
+
+statement ok
+insert into t1 values(1, 2), (2, 3), (3, 4);
+
+statement ok
+set inlist_to_join_threshold = 1;
+
+query T
+explain select * from t1 where a in (1, 2);
+----
+Exchange
+├── output columns: [t1.a (#0), t1.b (#1)]
+├── exchange type: Merge
+└── HashJoin
+    ├── output columns: [t1.a (#0), t1.b (#1)]
+    ├── join type: INNER
+    ├── build keys: [CAST(subquery_2 (#2) AS Int32 NULL)]
+    ├── probe keys: [t1.a (#0)]
+    ├── filters: []
+    ├── estimated rows: 2.00
+    ├── Exchange(Build)
+    │   ├── output columns: [col0 (#2)]
+    │   ├── exchange type: Broadcast
+    │   └── AggregateFinal
+    │       ├── output columns: [col0 (#2)]
+    │       ├── group by: [col0]
+    │       ├── aggregate functions: []
+    │       ├── estimated rows: 2.00
+    │       └── Exchange
+    │           ├── output columns: [col0 (#2)]
+    │           ├── exchange type: Hash(0)
+    │           └── AggregatePartial
+    │               ├── group by: [col0]
+    │               ├── aggregate functions: []
+    │               ├── estimated rows: 2.00
+    │               └── ConstantTableScan
+    │                   ├── output columns: [col0 (#2)]
+    │                   └── column 0: [1, 2]
+    └── TableScan(Probe)
+        ├── table: default.default.t1
+        ├── output columns: [a (#0), b (#1)]
+        ├── read rows: 3
+        ├── read bytes: 88
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [], limit: NONE]
+        └── estimated rows: 3.00
+
+
+statement ok
+drop table t1;
+
+statement ok
+unset inlist_to_join_threshold;
diff --git a/tests/sqllogictests/suites/mode/cluster/group_shuffle.test b/tests/sqllogictests/suites/mode/cluster/group_shuffle.test
@@ -45,8 +45,8 @@ Exchange
     ├── aggregate functions: []
     ├── estimated rows: 100000.00
     └── Exchange
-        ├── output columns: [#_group_by_key]
-        ├── exchange type: Hash(_group_by_key)
+        ├── output columns: [numbers_mt.number (#0)]
+        ├── exchange type: Hash(0)
         └── AggregatePartial
             ├── group by: [number]
             ├── aggregate functions: []