[ https://issues.apache.org/jira/browse/ARROW-15757?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17497462#comment-17497462 ]
Joris Van den Bossche commented on ARROW-15757: ----------------------------------------------- Indeed, we should probably ensure users can pass that keyword in write_to_dataset as well. Currently, the {{**kwargs}} are passed to the ParquetFileFormat write options (for parquet specific write options). Thanks for raising the issue! > [Python] Missing bindings for existing_data_behavior makes it impossible to > maintain old behavior > -------------------------------------------------------------------------------------------------- > > Key: ARROW-15757 > URL: https://issues.apache.org/jira/browse/ARROW-15757 > Project: Apache Arrow > Issue Type: Bug > Components: Parquet, Python > Affects Versions: 7.0.0 > Reporter: christophe bagot > Priority: Major > > Shouldn't the missing bindings reported earlier in > [https://github.com/apache/arrow/pull/11632] be propagated higher up [here in > the parquet.py > module|https://github.com/apache/arrow/blob/master/python/pyarrow/parquet.py#L2217]? > Passing **kwargs as is the case for {{write_table}} would do the trick I > think. > I am finding myself stuck while using pandas.to_parquet with > {{use_legacy_dataset=false}} and no way to set the {{existing_data_behavior}} > flag to {{overwrite_or_ignore}} > -- This message was sent by Atlassian Jira (v8.20.1#820001)