prashantwason opened a new pull request, #8493: URL: https://github.com/apache/hudi/pull/8493
- [HUDI-6098] Use bulk insert prepped for the initial write into MDT. ### Change Logs 1. Added a flag to HoodieTableMetadataWriter.commit to specify if the commit is an initial commit 2. For initial commit, bulkInsertPrepped API is used 3. Added a partitioner for MDT bulk insert which partitions the records based on their file Group. Since the records are already tagged before calling commit, this partitioner can retrieve the fileID and partition from the current location of the record. ### Impact Massive increase in read performance after initial creation of a index. Reduces the large read/write IO requirement for the first compaction in MDT. Reduces the duplicate storage of initial log files keeping the redundant initial commit data until cleaned. Faster initial commit as bulkInsert is more performant for billions of records than upsert which has a workload profiling stage. ### Risk level (write none, low medium or high below) None Already covered by existing unit tests. ### Documentation Update None ### Contributor's checklist - [ ] Read through [contributor's guide](https://hudi.apache.org/contribute/how-to-contribute) - [ ] Change Logs and Impact were stated clearly - [ ] Adequate tests were added if applicable - [ ] CI passed -- 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]
