This is an automated email from the ASF dual-hosted git repository.
yiguolei pushed a commit to branch dev-1.1.2
in repository https://gitbox.apache.org/repos/asf/doris.git
The following commit(s) were added to refs/heads/dev-1.1.2 by this push:
new 964fc7c21f [opt](join) speed up large data join (#11773)
964fc7c21f is described below
commit 964fc7c21f3e046cc286131d0ace7693e97085dc
Author: HappenLee <happen...@hotmail.com>
AuthorDate: Fri Aug 26 17:44:31 2022 +0800
[opt](join) speed up large data join (#11773)
Co-authored-by: lihaopeng <lihaop...@baidu.com>
---
be/src/vec/exec/join/vhash_join_node.cpp | 62 ++++++++++++++++++--------------
1 file changed, 36 insertions(+), 26 deletions(-)
diff --git a/be/src/vec/exec/join/vhash_join_node.cpp
b/be/src/vec/exec/join/vhash_join_node.cpp
index 98ffa85654..0e716c08b8 100644
--- a/be/src/vec/exec/join/vhash_join_node.cpp
+++ b/be/src/vec/exec/join/vhash_join_node.cpp
@@ -192,12 +192,7 @@ struct ProcessHashTableProbe {
if (_build_blocks.size() == 1) {
for (int i = 0; i < column_length; i++) {
if (output_slot_flags[i]) {
- auto column =
_build_blocks[0].get_by_position(i).column;
- if (mcol[i + column_offset]->is_nullable() xor
column->is_nullable()) {
- DCHECK(mcol[i + column_offset]->is_nullable() &&
- !column->is_nullable());
- column = make_nullable(column);
- }
+ auto& column =
_build_blocks[0].get_by_position(i).column;
mcol[i + column_offset]->insert_indices_from(
*column, _build_block_rows.data(),
_build_block_rows.data() + size);
} else {
@@ -205,6 +200,15 @@ struct ProcessHashTableProbe {
}
}
} else {
+ bool need_nullable_convert[column_length];
+ for (int i = 0; i < column_length; ++i) {
+ if (output_slot_flags[i]) {
+ need_nullable_convert[i] =
+ probe_all &&
+
!_build_blocks[0].get_by_position(i).column->is_nullable();
+ }
+ }
+
for (int i = 0; i < column_length; i++) {
if (output_slot_flags[i]) {
for (int j = 0; j < size; j++) {
@@ -214,17 +218,20 @@ struct ProcessHashTableProbe {
assert_cast<ColumnNullable*>(mcol[i +
column_offset].get())
->insert_join_null_data();
} else {
- auto column =
_build_blocks[_build_block_offsets[j]]
- .get_by_position(i)
- .column;
- if (mcol[i + column_offset]->is_nullable()
xor
- column->is_nullable()) {
- DCHECK(mcol[i +
column_offset]->is_nullable() &&
- !column->is_nullable());
- column = make_nullable(column);
+ auto& column =
_build_blocks[_build_block_offsets[j]]
+ .get_by_position(i)
+ .column;
+ if (need_nullable_convert[i]) {
+ assert_cast<ColumnNullable*>(mcol[i +
column_offset].get())
+ ->get_null_map_data()
+ .push_back(0);
+ assert_cast<ColumnNullable*>(mcol[i +
column_offset].get())
+ ->get_nested_column()
+ .insert_from(*column,
_build_block_rows[j]);
+ } else {
+ mcol[i +
column_offset]->insert_from(*column,
+
_build_block_rows[j]);
}
- mcol[i +
column_offset]->insert_from(*column,
-
_build_block_rows[j]);
}
} else {
if (_build_block_offsets[j] == -1) {
@@ -236,17 +243,20 @@ struct ProcessHashTableProbe {
// just insert default value
mcol[i + column_offset]->insert_default();
} else {
- auto column =
_build_blocks[_build_block_offsets[j]]
- .get_by_position(i)
- .column;
- if (mcol[i + column_offset]->is_nullable()
xor
- column->is_nullable()) {
- DCHECK(mcol[i +
column_offset]->is_nullable() &&
- !column->is_nullable());
- column = make_nullable(column);
+ auto& column =
_build_blocks[_build_block_offsets[j]]
+ .get_by_position(i)
+ .column;
+ if (need_nullable_convert[i]) {
+ assert_cast<ColumnNullable*>(mcol[i +
column_offset].get())
+ ->get_null_map_data()
+ .push_back(0);
+ assert_cast<ColumnNullable*>(mcol[i +
column_offset].get())
+ ->get_nested_column()
+ .insert_from(*column,
_build_block_rows[j]);
+ } else {
+ mcol[i +
column_offset]->insert_from(*column,
+
_build_block_rows[j]);
}
- mcol[i +
column_offset]->insert_from(*column,
-
_build_block_rows[j]);
}
}
}
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For additional commands, e-mail: commits-h...@doris.apache.org
