reader context had default implementation of get engine delete representation as null. The unmerged reader does not have handling for that. When compared with the other unmerged reader implementations they never process delete blocks, so the unmerged buffer is now updated to skip delete blocks
Impact
Spark npe is avoided and should be a slight perf boost. Also 1 less reader context method which is good.
Change Logs
reader context had default implementation of get engine delete representation as null. The unmerged reader does not have handling for that. When compared with the other unmerged reader implementations they never process delete blocks, so the unmerged buffer is now updated to skip delete blocks
Impact
Spark npe is avoided and should be a slight perf boost. Also 1 less reader context method which is good.
Risk level (write none, low medium or high below)
none
Documentation Update
N/A
Contributor's checklist