[ https://issues.apache.org/jira/browse/FLINK-19868?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17222700#comment-17222700 ]
Jark Wu commented on FLINK-19868: --------------------------------- I think {{csv.line-delimiter}} is not a good option for csv format, the {{line-delimiter}} should be an option of filesystem, because it is never used for message queues. For the row-wise formats, it should consume and produce the bytes for a row. The line splitting should be handled by connectors. I also searched the mailing list, and I didn't find any user asking questions about {{csv.line-delimiter}}, only some people are asking {{csv.field-delimiter}}. Therefore, I think we can remove {{csv.line-delimiter}} option (mention this in release note) and not append a new line in {{CsvRowDataSerializationSchema}}. What do you think? > Csv Serialization schema contains line delimiter > ------------------------------------------------ > > Key: FLINK-19868 > URL: https://issues.apache.org/jira/browse/FLINK-19868 > Project: Flink > Issue Type: Bug > Components: Formats (JSON, Avro, Parquet, ORC, SequenceFile) > Reporter: Jingsong Lee > Priority: Major > > CsvRowSerializationSchema.serialize(Row.of("f0", "f1")) => f0,f1\n > Csv Serialization schema is for one line, why contains line delimiter? -- This message was sent by Atlassian Jira (v8.3.4#803005)