github-actions[bot] commented on code in PR #62699:
URL: https://github.com/apache/doris/pull/62699#discussion_r3301269394
##########
be/src/exprs/function/function_tokenize.cpp:
##########
@@ -165,7 +166,13 @@ Status FunctionTokenize::execute_impl(FunctionContext*
/*context*/, Block& block
if (config.analyzer_name.empty() &&
config.parser_type == InvertedIndexParserType::PARSER_NONE) {
_do_tokenize_none(*col_left, dest_column_ptr);
- block.replace_by_position(result, std::move(dest_column_ptr));
+ if (left_const) {
+ block.replace_by_position(
Review Comment:
This behavior change needs regression coverage. The PR description uses
`SELECT tokenize('hello world', 'parser=english') FROM table_with_many_rows`,
but when both arguments are literals the generic
`PreparedFunctionImpl::default_implementation_for_constant_arguments` should
unwrap, execute one row, and wrap the result as a `ColumnConst` before this
`left_const` branch is reached. Please add a regression test that fails before
this patch and exercises the actual intended path, with multiple input rows and
deterministic expected output/row count, so we can verify this branch fixes a
real user-visible case rather than an untested lower-level edge case.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
