You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@doris.apache.org by "Baymine (via GitHub)" <gi...@apache.org> on 2026/05/08 07:10:49 UTC
Re: [PR] [fix](csv) Honor Hive CSV physical line boundaries [doris]
Baymine commented on code in PR #61614:
URL: https://github.com/apache/doris/pull/61614#discussion_r3206962511
##########
be/src/format/csv/csv_reader.cpp:
##########
@@ -585,6 +585,10 @@ Status CsvReader::_create_file_reader(bool need_schema) {
Status CsvReader::_create_line_reader() {
std::shared_ptr<TextLineReaderContextIf> text_line_reader_ctx;
+ // Hive TextInputFormat defines one record per physical line, so OpenCSVSerde must not
+ // merge rows across line delimiters even if a quote stays open.
+ const bool allow_multiline_records = !(_range.__isset.table_format_params &&
+ _range.table_format_params.table_format_type == "hive");
Review Comment:
This path isn't reachable for the CSV reader. table_format_type == "transactional_hive" is only set by `HiveScanNode.java:468` when HiveSplit.isACID() is true, and Hive ACID requires ORC storage by spec — ACID tables cannot be stored as TEXTFILE/CSV, so csv_reader.cpp::_create_line_reader() is never entered on that path. The parallel `new_json_reader.cpp:388` follows the same "hive"-only convention. I'd prefer to keep the check narrow rather than add code for an unreachable branch.
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org