database/doris
2
0
Fork 0
Code Issues Pull Requests Actions 3 Packages Projects Releases Wiki Activity

[Improvementation](join) empty_block shall be set true when build block only one row (#33721)

Browse Source 
empty_block shall be set true when build block only one row
This commit is contained in:
Pxl
2024-04-18 14:26:39 +08:00
committed by yiguolei
parent ea19224d14
commit f17ac173b4
3 changed files with 22 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
be/src/pipeline/exec/hashjoin_build_sink.cpp
View File

@ -155,21 +155,22 @@ bool HashJoinBuildSinkLocalState::build_unique() const {
void HashJoinBuildSinkLocalState::init_short_circuit_for_probe() {
auto& p = _parent->cast<HashJoinBuildSinkOperatorX>();
bool empty_block =
!_shared_state->build_block ||
!(_shared_state->build_block->rows() > 1); // build size always mock a row into block
_shared_state->short_circuit_for_probe =
(_shared_state->_has_null_in_build_side &&
p._join_op == TJoinOp::NULL_AWARE_LEFT_ANTI_JOIN && !p._is_mark_join) ||
(!_shared_state->build_block && p._join_op == TJoinOp::INNER_JOIN &&
!p._is_mark_join) ||
(!_shared_state->build_block && p._join_op == TJoinOp::LEFT_SEMI_JOIN &&
!p._is_mark_join) ||
(!_shared_state->build_block && p._join_op == TJoinOp::RIGHT_OUTER_JOIN) ||
(!_shared_state->build_block && p._join_op == TJoinOp::RIGHT_SEMI_JOIN) ||
(!_shared_state->build_block && p._join_op == TJoinOp::RIGHT_ANTI_JOIN);
(empty_block && p._join_op == TJoinOp::INNER_JOIN && !p._is_mark_join) ||
(empty_block && p._join_op == TJoinOp::LEFT_SEMI_JOIN && !p._is_mark_join) ||
(empty_block && p._join_op == TJoinOp::RIGHT_OUTER_JOIN) ||
(empty_block && p._join_op == TJoinOp::RIGHT_SEMI_JOIN) ||
(empty_block && p._join_op == TJoinOp::RIGHT_ANTI_JOIN);
//when build table rows is 0 and not have other_join_conjunct and not _is_mark_join and join type is one of LEFT_OUTER_JOIN/FULL_OUTER_JOIN/LEFT_ANTI_JOIN
//we could get the result is probe table + null-column(if need output)
_shared_state->empty_right_table_need_probe_dispose =
(!_shared_state->build_block && !p._have_other_join_conjunct && !p._is_mark_join) &&
(empty_block && !p._have_other_join_conjunct && !p._is_mark_join) &&
(p._join_op == TJoinOp::LEFT_OUTER_JOIN || p._join_op == TJoinOp::FULL_OUTER_JOIN ||
p._join_op == TJoinOp::LEFT_ANTI_JOIN);
}

17
be/src/pipeline/exec/hashjoin_probe_operator.cpp
View File

@ -247,7 +247,9 @@ Status HashJoinProbeOperatorX::pull(doris::RuntimeState* state, vectorized::Bloc
}
//TODO: this short circuit maybe could refactor, no need to check at here.
if (local_state._shared_state->empty_right_table_need_probe_dispose) {
// only support nereids
if (local_state._shared_state->empty_right_table_need_probe_dispose &&
!Base::_projections.empty()) {
// when build table rows is 0 and not have other_join_conjunct and join type is one of LEFT_OUTER_JOIN/FULL_OUTER_JOIN/LEFT_ANTI_JOIN
// we could get the result is probe table + null-column(if need output)
// If we use a short-circuit strategy, should return block directly by add additional null data.
@ -257,12 +259,6 @@ Status HashJoinProbeOperatorX::pull(doris::RuntimeState* state, vectorized::Bloc
return Status::OK();
}
vectorized::Block temp_block;
//get probe side output column
for (int i = 0; i < _left_output_slot_flags.size(); ++i) {
temp_block.insert(local_state._probe_block.get_by_position(i));
}
//create build side null column, if need output
for (int i = 0;
(_join_op != TJoinOp::LEFT_ANTI_JOIN) && i < _right_output_slot_flags.size(); ++i) {
@ -273,8 +269,8 @@ Status HashJoinProbeOperatorX::pull(doris::RuntimeState* state, vectorized::Bloc
vectorized::ColumnVector<vectorized::UInt8>::create(block_rows, 1);
auto nullable_column = vectorized::ColumnNullable::create(std::move(column),
std::move(null_map_column));
temp_block.insert({std::move(nullable_column), make_nullable(type),
_right_table_column_names[i]});
local_state._probe_block.insert({std::move(nullable_column), make_nullable(type),
_right_table_column_names[i]});
}
if (_is_outer_join) {
reinterpret_cast<vectorized::ColumnUInt8*>(
@ -290,8 +286,7 @@ Status HashJoinProbeOperatorX::pull(doris::RuntimeState* state, vectorized::Bloc
/// No need to check the block size in `_filter_data_and_build_output` because here dose not
/// increase the output rows count(just same as `_probe_block`'s rows count).
RETURN_IF_ERROR(local_state.filter_data_and_build_output(state, output_block, eos,
&temp_block, false));
temp_block.clear();
&local_state._probe_block, false));
local_state._probe_block.clear_column_data(_child_x->row_desc().num_materialized_slots());
return Status::OK();
}

12
be/src/vec/core/column_with_type_and_name.cpp
View File

@ -62,15 +62,17 @@ void ColumnWithTypeAndName::dump_structure(std::ostream& out) const {
out << name;
}
if (type)
if (type) {
out << " " << type->get_name();
else
} else {
out << " nullptr";
}
if (column)
out << ' ' << column->dump_structure();
else
if (column) {
out << ' ' << column->dump_structure() << "(use_count=" << column->use_count() << ')';
} else {
out << " nullptr";
}
}
String ColumnWithTypeAndName::dump_structure() const {