This is an automated email from the ASF dual-hosted git repository.
dataroaring pushed a commit to branch branch-3.0
in repository https://gitbox.apache.org/repos/asf/doris.git
The following commit(s) were added to refs/heads/branch-3.0 by this push:
new fa8df1b84db branch-3.0: [Fix](ShortCircuit) need to shrink char type
when read column store #50975 (#51332)
fa8df1b84db is described below
commit fa8df1b84db37e82efafc78e8eb1ece8c55ee7e3
Author: github-actions[bot]
<41898282+github-actions[bot]@users.noreply.github.com>
AuthorDate: Wed Jun 11 10:38:06 2025 +0800
branch-3.0: [Fix](ShortCircuit) need to shrink char type when read column
store #50975 (#51332)
Cherry-picked from #50975
---------
Co-authored-by: lihangyu <[email protected]>
---
be/src/olap/tablet_schema.cpp | 19 +++++++++++
be/src/olap/tablet_schema.h | 1 +
be/src/service/point_query_executor.cpp | 11 ++++--
be/test/olap/storage_types_test.cpp | 38 +++++++++++++++++++++
.../data/point_query_p0/test_point_query.out | Bin 9920 -> 9934 bytes
.../suites/point_query_p0/test_point_query.groovy | 21 +++++++++++-
6 files changed, 86 insertions(+), 4 deletions(-)
diff --git a/be/src/olap/tablet_schema.cpp b/be/src/olap/tablet_schema.cpp
index 947d6452b67..43a2a034722 100644
--- a/be/src/olap/tablet_schema.cpp
+++ b/be/src/olap/tablet_schema.cpp
@@ -37,6 +37,7 @@
#include "common/status.h"
#include "exec/tablet_info.h"
#include "olap/inverted_index_parser.h"
+#include "olap/olap_common.h"
#include "olap/olap_define.h"
#include "olap/tablet_column_object_pool.h"
#include "olap/types.h"
@@ -49,6 +50,8 @@
#include "vec/core/block.h"
#include "vec/data_types/data_type.h"
#include "vec/data_types/data_type_factory.hpp"
+#include "vec/data_types/data_type_map.h"
+#include "vec/data_types/data_type_struct.h"
#include "vec/json/path_in_data.h"
namespace doris {
@@ -460,6 +463,22 @@ uint32_t
TabletColumn::get_field_length_by_type(TPrimitiveType::type type, uint3
}
}
+bool TabletColumn::has_char_type() const {
+ switch (_type) {
+ case FieldType::OLAP_FIELD_TYPE_CHAR: {
+ return true;
+ }
+ case FieldType::OLAP_FIELD_TYPE_ARRAY:
+ case FieldType::OLAP_FIELD_TYPE_MAP:
+ case FieldType::OLAP_FIELD_TYPE_STRUCT: {
+ return std::any_of(_sub_columns.begin(), _sub_columns.end(),
+ [&](const auto& sub) -> bool { return
sub->has_char_type(); });
+ }
+ default:
+ return false;
+ }
+}
+
TabletColumn::TabletColumn() :
_aggregation(FieldAggregationMethod::OLAP_FIELD_AGGREGATION_NONE) {}
TabletColumn::TabletColumn(FieldAggregationMethod agg, FieldType type) {
diff --git a/be/src/olap/tablet_schema.h b/be/src/olap/tablet_schema.h
index 41eb8ccd1e3..ce114bf80f5 100644
--- a/be/src/olap/tablet_schema.h
+++ b/be/src/olap/tablet_schema.h
@@ -131,6 +131,7 @@ public:
int precision() const { return _precision; }
int frac() const { return _frac; }
inline bool visible() const { return _visible; }
+ bool has_char_type() const;
void set_aggregation_method(FieldAggregationMethod agg) {
_aggregation = agg;
diff --git a/be/src/service/point_query_executor.cpp
b/be/src/service/point_query_executor.cpp
index ee03495ffbd..73b34f2df7b 100644
--- a/be/src/service/point_query_executor.cpp
+++ b/be/src/service/point_query_executor.cpp
@@ -502,9 +502,14 @@ Status PointQueryExecutor::_lookup_row_data() {
vectorized::MutableColumnPtr column =
_result_block->get_by_position(pos).column->assume_mutable();
std::unique_ptr<ColumnIterator> iter;
- RETURN_IF_ERROR(segment->seek_and_read_by_rowid(
- *_tablet->tablet_schema(),
_reusable->tuple_desc()->slots()[pos], row_id,
- column, _read_stats, iter));
+ SlotDescriptor* slot = _reusable->tuple_desc()->slots()[pos];
+
RETURN_IF_ERROR(segment->seek_and_read_by_rowid(*_tablet->tablet_schema(), slot,
+ row_id,
column, _read_stats, iter));
+ if (_tablet->tablet_schema()
+ ->column_by_uid(slot->col_unique_id())
+ .has_char_type()) {
+ _result_block->get_by_position(pos).column =
column->get_shrinked_column();
+ }
}
}
}
diff --git a/be/test/olap/storage_types_test.cpp
b/be/test/olap/storage_types_test.cpp
index ab7b80a18f8..1eba8796c2d 100644
--- a/be/test/olap/storage_types_test.cpp
+++ b/be/test/olap/storage_types_test.cpp
@@ -239,4 +239,42 @@ TEST(ArrayTypeTest, copy_and_equal) {
CollectionValue(char_array, 3, null_signs));
}
+TEST(TypesTest, has_char_type) {
+ // Test basic types
+ TabletColumn
char_column(FieldAggregationMethod::OLAP_FIELD_AGGREGATION_NONE,
+ FieldType::OLAP_FIELD_TYPE_CHAR);
+ EXPECT_TRUE(char_column.has_char_type());
+
+ TabletColumn
int_column(FieldAggregationMethod::OLAP_FIELD_AGGREGATION_NONE,
+ FieldType::OLAP_FIELD_TYPE_INT);
+ EXPECT_FALSE(int_column.has_char_type());
+
+ // Test array type with char element
+ TabletColumn
array_column(FieldAggregationMethod::OLAP_FIELD_AGGREGATION_NONE,
+ FieldType::OLAP_FIELD_TYPE_ARRAY);
+ TabletColumn
array_element(FieldAggregationMethod::OLAP_FIELD_AGGREGATION_NONE,
+ FieldType::OLAP_FIELD_TYPE_CHAR);
+ array_column.add_sub_column(array_element);
+ EXPECT_TRUE(array_column.has_char_type());
+
+ // Test array type with non-char element
+ TabletColumn
array_column2(FieldAggregationMethod::OLAP_FIELD_AGGREGATION_NONE,
+ FieldType::OLAP_FIELD_TYPE_ARRAY);
+ TabletColumn
array_element2(FieldAggregationMethod::OLAP_FIELD_AGGREGATION_NONE,
+ FieldType::OLAP_FIELD_TYPE_INT);
+ array_column2.add_sub_column(array_element2);
+ EXPECT_FALSE(array_column2.has_char_type());
+
+ // Test nested array with char element
+ TabletColumn
nested_array(FieldAggregationMethod::OLAP_FIELD_AGGREGATION_NONE,
+ FieldType::OLAP_FIELD_TYPE_ARRAY);
+ TabletColumn
inner_array(FieldAggregationMethod::OLAP_FIELD_AGGREGATION_NONE,
+ FieldType::OLAP_FIELD_TYPE_ARRAY);
+ TabletColumn
char_element(FieldAggregationMethod::OLAP_FIELD_AGGREGATION_NONE,
+ FieldType::OLAP_FIELD_TYPE_CHAR);
+ inner_array.add_sub_column(char_element);
+ nested_array.add_sub_column(inner_array);
+ EXPECT_TRUE(nested_array.has_char_type());
+}
+
} // namespace doris
diff --git a/regression-test/data/point_query_p0/test_point_query.out
b/regression-test/data/point_query_p0/test_point_query.out
index 75822c3cf88..8572ed52a68 100644
Binary files a/regression-test/data/point_query_p0/test_point_query.out and
b/regression-test/data/point_query_p0/test_point_query.out differ
diff --git a/regression-test/suites/point_query_p0/test_point_query.groovy
b/regression-test/suites/point_query_p0/test_point_query.groovy
index f2b22064d9c..fbf96d9f478 100644
--- a/regression-test/suites/point_query_p0/test_point_query.groovy
+++ b/regression-test/suites/point_query_p0/test_point_query.groovy
@@ -425,4 +425,23 @@ suite("test_point_query", "nonConcurrent") {
assertNotEquals(result1, result2)
}
}
-}
\ No newline at end of file
+ // test shrink char type
+ sql "DROP TABLE IF EXISTS table_with_chars"
+ sql """
+ CREATE TABLE `table_with_chars` (
+ `col1` smallint NOT NULL,
+ `col2` int NOT NULL,
+ `loc3` char(10) NOT NULL,
+ `value` char(10) NOT NULL,
+ INDEX col3 (`loc3`) USING INVERTED,
+ INDEX col2 (`col2`) USING INVERTED )
+ ENGINE=OLAP UNIQUE KEY(`col1`)
+ DISTRIBUTED BY HASH(`col1`) BUCKETS 1
+ PROPERTIES ( "replication_allocation" = "tag.location.default: 1",
"bloom_filter_columns" = "col1", "row_store_columns" = "col1",
"enable_mow_light_delete" = "false" );
+ """
+ sql "insert into table_with_chars values (-10, 20, 'aabc', 'value')"
+ sql "insert into table_with_chars values (10, 20, 'aabc', 'value');"
+ sql "insert into table_with_chars values (20, 30, 'aabc', 'value');"
+ sql "set enable_short_circuit_query = true"
+ qt_sql "select length(loc3) from table_with_chars where col1 = 10"
+}
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
