[Enhancement] enable adaptive partition hash join by default (StarRoc…

…ks#51224) Signed-off-by: stdpain <drfeng08@gmail.com>
iampfc · Sep 27, 2024 · 26de537 · 26de537
1 parent 65f3007
commit 26de537
Show file tree

Hide file tree

Showing 4 changed files with 14 additions and 7 deletions.
diff --git a/be/src/exec/hash_join_components.cpp b/be/src/exec/hash_join_components.cpp
@@ -627,7 +627,7 @@ int64_t AdaptivePartitionHashJoinBuilder::ht_mem_usage() const {
 
 Status AdaptivePartitionHashJoinBuilder::_convert_to_single_partition() {
     // merge all partition data to the first partition
-    for (size_t i = 1; i < _partition_num; ++i) {
+    for (size_t i = 1; i < _builders.size(); ++i) {
         _builders[0]->hash_table().merge_ht(_builders[i]->hash_table());
     }
     _builders.resize(1);

diff --git a/be/src/exec/pipeline/hashjoin/spillable_hash_join_build_operator.cpp b/be/src/exec/pipeline/hashjoin/spillable_hash_join_build_operator.cpp
@@ -249,13 +249,18 @@ StatusOr<std::function<StatusOr<ChunkPtr>()>> SpillableHashJoinBuildOperator::_c
 
     return [this]() -> StatusOr<ChunkPtr> {
         if (_hash_table_build_chunk_slice.empty()) {
-            if (_hash_table_iterate_idx + 1 >= _hash_tables.size()) {
+            _hash_table_iterate_idx++;
+            for (; _hash_table_iterate_idx < _hash_tables.size(); _hash_table_iterate_idx++) {
+                auto build_chunk = _hash_tables[_hash_table_iterate_idx]->get_build_chunk();
+                if (build_chunk->num_rows() > 0) {
+                    _hash_table_build_chunk_slice.reset(build_chunk);
+                    _hash_table_build_chunk_slice.skip(kHashJoinKeyColumnOffset);
+                    break;
+                }
+            }
+            if (_hash_table_build_chunk_slice.empty()) {
                 _join_builder->hash_join_builder()->reset(_join_builder->hash_table_param());
                 return Status::EndOfFile("eos");
-            } else {
-                _hash_table_iterate_idx++;
-                _hash_table_build_chunk_slice.reset(_hash_tables[_hash_table_iterate_idx]->get_build_chunk());
-                _hash_table_build_chunk_slice.skip(kHashJoinKeyColumnOffset);
             }
         }
 

diff --git a/be/src/types/logical_type_infra.h b/be/src/types/logical_type_infra.h
@@ -149,6 +149,8 @@ auto type_dispatch_all(LogicalType ltype, Functor fun, Args... args) {
     switch (ltype) {
         APPLY_FOR_ALL_SCALAR_TYPE_WITH_NULL(_TYPE_DISPATCH_CASE)
         _TYPE_DISPATCH_CASE(TYPE_ARRAY)
+        _TYPE_DISPATCH_CASE(TYPE_STRUCT)
+        _TYPE_DISPATCH_CASE(TYPE_MAP)
         _TYPE_DISPATCH_CASE(TYPE_HLL)
         _TYPE_DISPATCH_CASE(TYPE_OBJECT)
         _TYPE_DISPATCH_CASE(TYPE_PERCENTILE)

diff --git a/fe/fe-core/src/main/java/com/starrocks/qe/SessionVariable.java b/fe/fe-core/src/main/java/com/starrocks/qe/SessionVariable.java
@@ -1263,7 +1263,7 @@ public static MaterializedViewRewriteMode parse(String str) {
     private boolean joinLateMaterialization = false;
 
     @VariableMgr.VarAttr(name = ENABLE_PARTITION_HASH_JOIN)
-    private boolean enablePartitionHashJoin = false;
+    private boolean enablePartitionHashJoin = true;
 
     @VariableMgr.VarAttr(name = ENABLE_PRUNE_COLUMN_AFTER_INDEX_FILTER, flag = VariableMgr.INVISIBLE)
     private boolean enablePruneColumnAfterIndexFilter = true;