[ https://issues.apache.org/jira/browse/ARROW-9332?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17156646#comment-17156646 ]
Joris Van den Bossche commented on ARROW-9332: ---------------------------------------------- Basic pickling of RowGroupInfo (that at least preserves the row group id) was added in ARROW-9321 (https://github.com/apache/arrow/pull/7692). Further enhancement could also preserve num_rows, total_byte_size and statistics (but that is less a priority). Pickling the statistics will first require pickling of scalars -> ARROW-9394 > [Python][Dataset] Support pickling of ParquetFileFragment's RowGroupInfo > ------------------------------------------------------------------------ > > Key: ARROW-9332 > URL: https://issues.apache.org/jira/browse/ARROW-9332 > Project: Apache Arrow > Issue Type: Improvement > Components: Python > Reporter: Joris Van den Bossche > Priority: Major > Labels: dataset, dataset-dask-integration > > Follow-up on ARROW-8651 to ensure we can also preserve the statistics > information of {{RowGroupInfo}} of a {{ParquetFileFragment}} -- This message was sent by Atlassian Jira (v8.3.4#803005)