This is an automated email from the ASF dual-hosted git repository.
morningman pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/doris.git
The following commit(s) were added to refs/heads/master by this push:
new ce595c27c54 [fix](serde)fix the bug in
DataTypeNullableSerDe.deserialize_column_from_fixed_json (#41217)
ce595c27c54 is described below
commit ce595c27c54985b9b9c29768eed8593ad64c0c01
Author: daidai <[email protected]>
AuthorDate: Fri Oct 11 18:19:53 2024 +0800
[fix](serde)fix the bug in
DataTypeNullableSerDe.deserialize_column_from_fixed_json (#41217)
---
be/src/vec/data_types/serde/data_type_datetimev2_serde.cpp | 6 ++++++
be/src/vec/data_types/serde/data_type_datev2_serde.cpp | 6 ++++++
be/src/vec/data_types/serde/data_type_decimal_serde.cpp | 6 ++++++
be/src/vec/data_types/serde/data_type_nullable_serde.cpp | 5 ++++-
be/src/vec/data_types/serde/data_type_number_serde.cpp | 6 ++++++
be/src/vec/data_types/serde/data_type_serde.h | 11 ++++++++++-
be/src/vec/data_types/serde/data_type_string_serde.h | 6 ++++++
be/src/vec/exec/format/parquet/vparquet_group_reader.cpp | 1 +
8 files changed, 45 insertions(+), 2 deletions(-)
diff --git a/be/src/vec/data_types/serde/data_type_datetimev2_serde.cpp
b/be/src/vec/data_types/serde/data_type_datetimev2_serde.cpp
index e57af914d43..e8238af4eee 100644
--- a/be/src/vec/data_types/serde/data_type_datetimev2_serde.cpp
+++ b/be/src/vec/data_types/serde/data_type_datetimev2_serde.cpp
@@ -252,6 +252,9 @@ Status DataTypeDateTimeV2SerDe::write_column_to_orc(const
std::string& timezone,
Status DataTypeDateTimeV2SerDe::deserialize_column_from_fixed_json(
IColumn& column, Slice& slice, int rows, int* num_deserialized,
const FormatOptions& options) const {
+ if (rows < 1) [[unlikely]] {
+ return Status::OK();
+ }
Status st = deserialize_one_cell_from_json(column, slice, options);
if (!st.ok()) {
return st;
@@ -264,6 +267,9 @@ Status
DataTypeDateTimeV2SerDe::deserialize_column_from_fixed_json(
void DataTypeDateTimeV2SerDe::insert_column_last_value_multiple_times(IColumn&
column,
int
times) const {
+ if (times < 1) [[unlikely]] {
+ return;
+ }
auto& col = static_cast<ColumnVector<UInt64>&>(column);
auto sz = col.size();
UInt64 val = col.get_element(sz - 1);
diff --git a/be/src/vec/data_types/serde/data_type_datev2_serde.cpp
b/be/src/vec/data_types/serde/data_type_datev2_serde.cpp
index f2d595b87c4..95109ee408c 100644
--- a/be/src/vec/data_types/serde/data_type_datev2_serde.cpp
+++ b/be/src/vec/data_types/serde/data_type_datev2_serde.cpp
@@ -178,6 +178,9 @@ Status DataTypeDateV2SerDe::write_column_to_orc(const
std::string& timezone, con
Status DataTypeDateV2SerDe::deserialize_column_from_fixed_json(IColumn&
column, Slice& slice,
int rows, int*
num_deserialized,
const
FormatOptions& options) const {
+ if (rows < 1) [[unlikely]] {
+ return Status::OK();
+ }
Status st = deserialize_one_cell_from_json(column, slice, options);
if (!st.ok()) {
return st;
@@ -189,6 +192,9 @@ Status
DataTypeDateV2SerDe::deserialize_column_from_fixed_json(IColumn& column,
void DataTypeDateV2SerDe::insert_column_last_value_multiple_times(IColumn&
column,
int times)
const {
+ if (times < 1) [[unlikely]] {
+ return;
+ }
auto& col = static_cast<ColumnVector<UInt32>&>(column);
auto sz = col.size();
UInt32 val = col.get_element(sz - 1);
diff --git a/be/src/vec/data_types/serde/data_type_decimal_serde.cpp
b/be/src/vec/data_types/serde/data_type_decimal_serde.cpp
index e979211d6d7..acb09ee773e 100644
--- a/be/src/vec/data_types/serde/data_type_decimal_serde.cpp
+++ b/be/src/vec/data_types/serde/data_type_decimal_serde.cpp
@@ -280,6 +280,9 @@ template <typename T>
Status DataTypeDecimalSerDe<T>::deserialize_column_from_fixed_json(
IColumn& column, Slice& slice, int rows, int* num_deserialized,
const FormatOptions& options) const {
+ if (rows < 1) [[unlikely]] {
+ return Status::OK();
+ }
Status st = deserialize_one_cell_from_json(column, slice, options);
if (!st.ok()) {
return st;
@@ -293,6 +296,9 @@ Status
DataTypeDecimalSerDe<T>::deserialize_column_from_fixed_json(
template <typename T>
void DataTypeDecimalSerDe<T>::insert_column_last_value_multiple_times(IColumn&
column,
int
times) const {
+ if (times < 1) [[unlikely]] {
+ return;
+ }
auto& col = static_cast<ColumnDecimal<T>&>(column);
auto sz = col.size();
diff --git a/be/src/vec/data_types/serde/data_type_nullable_serde.cpp
b/be/src/vec/data_types/serde/data_type_nullable_serde.cpp
index 1af85bd040d..3b46e0e784f 100644
--- a/be/src/vec/data_types/serde/data_type_nullable_serde.cpp
+++ b/be/src/vec/data_types/serde/data_type_nullable_serde.cpp
@@ -131,12 +131,15 @@ Status
DataTypeNullableSerDe::deserialize_column_from_hive_text_vector(
Status DataTypeNullableSerDe::deserialize_column_from_fixed_json(
IColumn& column, Slice& slice, int rows, int* num_deserialized,
const FormatOptions& options) const {
+ if (rows < 1) [[unlikely]] {
+ return Status::OK();
+ }
auto& col = static_cast<ColumnNullable&>(column);
Status st = deserialize_one_cell_from_json(column, slice, options);
if (!st.ok()) {
return st;
}
- if (rows - 1 != 0) {
+ if (rows > 1) {
auto& null_map = col.get_null_map_data();
auto& nested_column = col.get_nested_column();
diff --git a/be/src/vec/data_types/serde/data_type_number_serde.cpp
b/be/src/vec/data_types/serde/data_type_number_serde.cpp
index 299779ea267..efa41e346bf 100644
--- a/be/src/vec/data_types/serde/data_type_number_serde.cpp
+++ b/be/src/vec/data_types/serde/data_type_number_serde.cpp
@@ -228,6 +228,9 @@ template <typename T>
Status DataTypeNumberSerDe<T>::deserialize_column_from_fixed_json(
IColumn& column, Slice& slice, int rows, int* num_deserialized,
const FormatOptions& options) const {
+ if (rows < 1) [[unlikely]] {
+ return Status::OK();
+ }
Status st = deserialize_one_cell_from_json(column, slice, options);
if (!st.ok()) {
return st;
@@ -241,6 +244,9 @@ Status
DataTypeNumberSerDe<T>::deserialize_column_from_fixed_json(
template <typename T>
void DataTypeNumberSerDe<T>::insert_column_last_value_multiple_times(IColumn&
column,
int
times) const {
+ if (times < 1) [[unlikely]] {
+ return;
+ }
auto& col = static_cast<ColumnVector<T>&>(column);
auto sz = col.size();
T val = col.get_element(sz - 1);
diff --git a/be/src/vec/data_types/serde/data_type_serde.h
b/be/src/vec/data_types/serde/data_type_serde.h
index 262f9cae6a8..46236faa926 100644
--- a/be/src/vec/data_types/serde/data_type_serde.h
+++ b/be/src/vec/data_types/serde/data_type_serde.h
@@ -243,17 +243,26 @@ public:
virtual Status deserialize_column_from_fixed_json(IColumn& column, Slice&
slice, int rows,
int* num_deserialized,
const FormatOptions&
options) const {
+ //In this function implementation, we need to consider the case where
rows is 0, 1, and other larger integers.
+ if (rows < 1) [[unlikely]] {
+ return Status::OK();
+ }
Status st = deserialize_one_cell_from_json(column, slice, options);
if (!st.ok()) {
*num_deserialized = 0;
return st;
}
- insert_column_last_value_multiple_times(column, rows - 1);
+ if (rows > 1) [[likely]] {
+ insert_column_last_value_multiple_times(column, rows - 1);
+ }
*num_deserialized = rows;
return Status::OK();
}
// Insert the last value to the end of this column multiple times.
virtual void insert_column_last_value_multiple_times(IColumn& column, int
times) const {
+ if (times < 1) [[unlikely]] {
+ return;
+ }
//If you try to simplify this operation by using
`column.insert_many_from(column, column.size() - 1, rows - 1);`
// you are likely to get incorrect data results.
MutableColumnPtr dum_col = column.clone_empty();
diff --git a/be/src/vec/data_types/serde/data_type_string_serde.h
b/be/src/vec/data_types/serde/data_type_string_serde.h
index fe09ff615f4..583772c5825 100644
--- a/be/src/vec/data_types/serde/data_type_string_serde.h
+++ b/be/src/vec/data_types/serde/data_type_string_serde.h
@@ -218,6 +218,9 @@ public:
Status deserialize_column_from_fixed_json(IColumn& column, Slice& slice,
int rows,
int* num_deserialized,
const FormatOptions& options)
const override {
+ if (rows < 1) [[unlikely]] {
+ return Status::OK();
+ }
Status st = deserialize_one_cell_from_json(column, slice, options);
if (!st.ok()) {
return st;
@@ -229,6 +232,9 @@ public:
}
void insert_column_last_value_multiple_times(IColumn& column, int times)
const override {
+ if (times < 1) [[unlikely]] {
+ return;
+ }
auto& col = static_cast<ColumnString&>(column);
auto sz = col.size();
diff --git a/be/src/vec/exec/format/parquet/vparquet_group_reader.cpp
b/be/src/vec/exec/format/parquet/vparquet_group_reader.cpp
index 08ecb601f39..37e82774c39 100644
--- a/be/src/vec/exec/format/parquet/vparquet_group_reader.cpp
+++ b/be/src/vec/exec/format/parquet/vparquet_group_reader.cpp
@@ -654,6 +654,7 @@ Status RowGroupReader::_fill_partition_columns(
auto _text_serde = slot_desc->get_data_type_ptr()->get_serde();
Slice slice(value.data(), value.size());
int num_deserialized = 0;
+ // Be careful when reading empty rows from parquet row groups.
if (_text_serde->deserialize_column_from_fixed_json(*col_ptr, slice,
rows,
&num_deserialized,
_text_formatOptions) != Status::OK()) {
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]