This is an automated email from the ASF dual-hosted git repository.
xuyang pushed a commit to branch branch-2.0
in repository https://gitbox.apache.org/repos/asf/doris.git
The following commit(s) were added to refs/heads/branch-2.0 by this push:
new 9beaab24ecc [fix](branch20) fix unexpected be core in string search
function (#31460)
9beaab24ecc is described below
commit 9beaab24ecc41592925e029188fda73aa987384c
Author: xy720 <[email protected]>
AuthorDate: Tue Feb 27 17:29:22 2024 +0800
[fix](branch20) fix unexpected be core in string search function (#31460)
cherry-pick: #31312
---
.../functions/functions_multi_string_position.cpp | 15 +++++++------
.../functions/functions_multi_string_search.cpp | 15 +++++++------
.../apache/doris/analysis/FunctionCallExpr.java | 3 +++
.../java/org/apache/doris/catalog/FunctionSet.java | 6 +++++
.../test_multi_string_position.groovy | 24 ++++++++++++++++++++
.../test_multi_string_search.groovy | 26 +++++++++++++++++++++-
6 files changed, 74 insertions(+), 15 deletions(-)
diff --git a/be/src/vec/functions/functions_multi_string_position.cpp
b/be/src/vec/functions/functions_multi_string_position.cpp
index f3230325638..47597145913 100644
--- a/be/src/vec/functions/functions_multi_string_position.cpp
+++ b/be/src/vec/functions/functions_multi_string_position.cpp
@@ -80,26 +80,22 @@ public:
Status execute_impl(FunctionContext* context, Block& block, const
ColumnNumbers& arguments,
size_t result, size_t input_rows_count) override {
auto haystack_column = block.get_by_position(arguments[0]).column;
- auto haystack_ptr = haystack_column;
-
auto needles_column = block.get_by_position(arguments[1]).column;
- auto needles_ptr = needles_column;
bool haystack_nullable = false;
bool needles_nullable = false;
if (haystack_column->is_nullable()) {
- haystack_ptr =
check_and_get_column<ColumnNullable>(haystack_column.get())
- ->get_nested_column_ptr();
haystack_nullable = true;
}
if (needles_column->is_nullable()) {
- needles_ptr =
check_and_get_column<ColumnNullable>(needles_column.get())
- ->get_nested_column_ptr();
needles_nullable = true;
}
+ auto haystack_ptr = remove_nullable(haystack_column);
+ auto needles_ptr = remove_nullable(needles_column);
+
const ColumnString* col_haystack_vector =
check_and_get_column<ColumnString>(&*haystack_ptr);
const ColumnConst* col_haystack_const =
@@ -110,6 +106,11 @@ public:
const ColumnConst* col_needles_const =
check_and_get_column_const<ColumnArray>(needles_ptr.get());
+ if (!col_needles_const && !col_needles_vector)
+ return Status::InvalidArgument(
+ "function '{}' encountered unsupported needles column,
found {}", name,
+ needles_column->get_name());
+
if (col_haystack_const && col_needles_vector) {
return Status::InvalidArgument(
"function '{}' doesn't support search with non-constant
needles "
diff --git a/be/src/vec/functions/functions_multi_string_search.cpp
b/be/src/vec/functions/functions_multi_string_search.cpp
index bc7b68dea59..bcd5ad0a5d5 100644
--- a/be/src/vec/functions/functions_multi_string_search.cpp
+++ b/be/src/vec/functions/functions_multi_string_search.cpp
@@ -78,26 +78,22 @@ public:
Status execute_impl(FunctionContext* context, Block& block, const
ColumnNumbers& arguments,
size_t result, size_t input_rows_count) override {
auto haystack_column = block.get_by_position(arguments[0]).column;
- auto haystack_ptr = haystack_column;
-
auto needles_column = block.get_by_position(arguments[1]).column;
- auto needles_ptr = needles_column;
bool haystack_nullable = false;
bool needles_nullable = false;
if (haystack_column->is_nullable()) {
- haystack_ptr =
check_and_get_column<ColumnNullable>(haystack_column.get())
- ->get_nested_column_ptr();
haystack_nullable = true;
}
if (needles_column->is_nullable()) {
- needles_ptr =
check_and_get_column<ColumnNullable>(needles_column.get())
- ->get_nested_column_ptr();
needles_nullable = true;
}
+ auto haystack_ptr = remove_nullable(haystack_column);
+ auto needles_ptr = remove_nullable(needles_column);
+
const ColumnString* col_haystack_vector =
check_and_get_column<ColumnString>(&*haystack_ptr);
const ColumnConst* col_haystack_const =
@@ -108,6 +104,11 @@ public:
const ColumnConst* col_needles_const =
check_and_get_column_const<ColumnArray>(needles_ptr.get());
+ if (!col_needles_const && !col_needles_vector)
+ return Status::InvalidArgument(
+ "function '{}' encountered unsupported needles column,
found {}", name,
+ needles_column->get_name());
+
if (col_haystack_const && col_needles_vector)
return Status::InvalidArgument(
"function '{}' doesn't support search with non-constant
needles "
diff --git
a/fe/fe-core/src/main/java/org/apache/doris/analysis/FunctionCallExpr.java
b/fe/fe-core/src/main/java/org/apache/doris/analysis/FunctionCallExpr.java
index 9f607ca7311..2e8f443f6ba 100644
--- a/fe/fe-core/src/main/java/org/apache/doris/analysis/FunctionCallExpr.java
+++ b/fe/fe-core/src/main/java/org/apache/doris/analysis/FunctionCallExpr.java
@@ -81,6 +81,9 @@ public class FunctionCallExpr extends Expr {
String.CASE_INSENSITIVE_ORDER)
.add("round").add("round_bankers").add("ceil").add("floor")
.add("truncate").add("dround").add("dceil").add("dfloor").build();
+ public static final ImmutableSet<String> STRING_SEARCH_FUNCTION_SET = new
ImmutableSortedSet.Builder(
+ String.CASE_INSENSITIVE_ORDER)
+ .add("multi_search_all_positions").add("multi_match_any").build();
private final AtomicBoolean addOnce = new AtomicBoolean(false);
diff --git a/fe/fe-core/src/main/java/org/apache/doris/catalog/FunctionSet.java
b/fe/fe-core/src/main/java/org/apache/doris/catalog/FunctionSet.java
index dc0a43c6265..99ab159f25d 100644
--- a/fe/fe-core/src/main/java/org/apache/doris/catalog/FunctionSet.java
+++ b/fe/fe-core/src/main/java/org/apache/doris/catalog/FunctionSet.java
@@ -501,6 +501,12 @@ public class FunctionSet<T> {
return false;
}
}
+ if
(FunctionCallExpr.STRING_SEARCH_FUNCTION_SET.contains(desc.functionName())) {
+ if (descArgTypes[1].isStringType() &&
candicateArgTypes[1].isArrayType()) {
+ // The needles arg of search functions should not be allowed
to cast from string.
+ return false;
+ }
+ }
// If set `roundPreciseDecimalV2Value`, only use decimalv3 as target
type to execute round function
if (ConnectContext.get() != null
&&
ConnectContext.get().getSessionVariable().roundPreciseDecimalV2Value
diff --git
a/regression-test/suites/query_p0/sql_functions/search_functions/test_multi_string_position.groovy
b/regression-test/suites/query_p0/sql_functions/search_functions/test_multi_string_position.groovy
index f6a8aa110e9..1c82fc7c6e9 100644
---
a/regression-test/suites/query_p0/sql_functions/search_functions/test_multi_string_position.groovy
+++
b/regression-test/suites/query_p0/sql_functions/search_functions/test_multi_string_position.groovy
@@ -54,4 +54,28 @@ suite("test_multi_string_position") {
qt_select3 "select
multi_search_all_positions('mpswgtljbbrmivkcglamemayfn', ['', 'm',
'saejhpnfgfq', 'rzanrkdssmmkanqjpfi', 'oputeneprgoowg', 'mp', '', '',
'wgtljbbrmivkcglamemay', 'cbpthtrgrmgfypizi', 'tl', 'tlj', 'xuhs',
'brmivkcglamemayfn', '', 'gtljb'])"
qt_select4 "select
multi_search_all_positions('arbphzbbecypbzsqsljurtddve', ['arbphzb',
'mnrboimjfijnti', 'cikcrd', 'becypbz', 'z',
'uocmqgnczhdcrvtqrnaxdxjjlhakoszuwc', 'bbe', '', 'bp', 'yhltnexlpdijkdzt',
'jkwjmrckvgmccmmrolqvy', 'vdxmicjmfbtsbqqmqcgtnrvdgaucsgspwg',
'witlfqwvhmmyjrnrzttrikhhsrd', 'pbzsqsljurt'])"
qt_select5 "select
multi_search_all_positions('aizovxqpzcbbxuhwtiaaqhdqjdei', ['qpzcbbxuhw',
'jugrpglqbm', 'dspwhzpyjohhtizegrnswhjfpdz', 'pzcbbxuh', 'vayzeszlycke', 'i',
'gvrontcpqavsjxtjwzgwxugiyhkhmhq', 'gyzmeroxztgaurmrqwtmsxcqnxaezuoapatvu',
'xqpzc', 'mjiswsvlvlpqrhhptqq', 'iz', 'hmzjxxfjsvcvdpqwtrdrp',
'zovxqpzcbbxuhwtia', 'ai'])"
+
+ try {
+ sql "select multi_search_all_positions(content, 'hello') from
${table_name} order by col1"
+ } catch (Exception ex) {
+ assert("${ex}".contains("errCode = 2, detailMessage = No matching
function with signature: multi_search_all_positions"))
+ }
+
+ try {
+ sql "select multi_search_all_positions(content, 'hello, !, world,
Hello, World') from ${table_name} order by col1"
+ } catch (Exception ex) {
+ assert("${ex}".contains("errCode = 2, detailMessage = No matching
function with signature: multi_search_all_positions"))
+ }
+
+ try {
+ sql "select multi_search_all_positions(content, '[hello]') from
${table_name} order by col1"
+ } catch (Exception ex) {
+ assert("${ex}".contains("errCode = 2, detailMessage = No matching
function with signature: multi_search_all_positions"))
+ }
+
+ try {
+ sql "select multi_search_all_positions(content, '[hello, !, world,
Hello, World]') from ${table_name} order by col1"
+ } catch (Exception ex) {
+ assert("${ex}".contains("errCode = 2, detailMessage = No matching
function with signature: multi_search_all_positions"))
+ }
}
diff --git
a/regression-test/suites/query_p0/sql_functions/search_functions/test_multi_string_search.groovy
b/regression-test/suites/query_p0/sql_functions/search_functions/test_multi_string_search.groovy
index 5a3229ce361..3a0a9b061fe 100644
---
a/regression-test/suites/query_p0/sql_functions/search_functions/test_multi_string_search.groovy
+++
b/regression-test/suites/query_p0/sql_functions/search_functions/test_multi_string_search.groovy
@@ -71,4 +71,28 @@ suite("test_multi_string_search") {
qt_select "select multi_match_any('ldrzgttlqaphekkkdukgngl',
['gttlqaphekkkdukgn', 'ekkkd', 'gttlqaphe', 'qaphek', 'h', 'kdu', 'he', 'phek',
'', 'drzgttlqaphekkkd'])"
qt_select "select multi_match_any('ololo', ['ololo', 'ololo', 'ololo'])"
qt_select "select multi_match_any('khljxzxlpcrxpkrfybbfk', ['k'])"
-}
\ No newline at end of file
+
+ try {
+ sql "select multi_match_any(content, 'hello') from ${table_name} order
by col1"
+ } catch (Exception ex) {
+ assert("${ex}".contains("errCode = 2, detailMessage = No matching
function with signature: multi_match_any"))
+ }
+
+ try {
+ sql "select multi_match_any(content, 'hello, !, world, Hello, World')
from ${table_name} order by col1"
+ } catch (Exception ex) {
+ assert("${ex}".contains("errCode = 2, detailMessage = No matching
function with signature: multi_match_any"))
+ }
+
+ try {
+ sql "select multi_match_any(content, '[hello]') from ${table_name}
order by col1"
+ } catch (Exception ex) {
+ assert("${ex}".contains("errCode = 2, detailMessage = No matching
function with signature: multi_match_any"))
+ }
+
+ try {
+ sql "select multi_match_any(content, '[hello, !, world, Hello,
World]') from ${table_name} order by col1"
+ } catch (Exception ex) {
+ assert("${ex}".contains("errCode = 2, detailMessage = No matching
function with signature: multi_match_any"))
+ }
+}
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
