I think SQLConf doesn't need defs anymore. In the beginning, SQLConf lived
in sql/core, so we have to add defs if the code in sql/catalyst needs to
access configs. Now SQLConf is in sql/catalyst (this was done a few years
ago), defs are only needed if we have some special logic that is not just
This is correct. It's true by default so that AQE doesn't have performance
regression. If you run a benchmark, larger parallelism usually means better
performance. However, it's recommended to set it to false, so that AQE can
give better resource utilization, which is good for a busy Spark
Sharing some things I learned looking into the Delta Lake issue:
- This was a read after write inconsistency _all on the driver_.
Specifically it currently uses the FileSystem API for reading table logs
for greater compatibility, but the FileContext API for writes for atomic
renames. This led to