Github user marmbrus commented on the pull request:
https://github.com/apache/spark/pull/8922#issuecomment-145632836
"daily sql query" is not sufficiently descriptive. Please post actual
benchmark results with code when making pull requests that claim to improve
performance. It would also be good to evaluate the cost in degenerate cases.
For example, I think you are adding an object allocation per input tuple when
boxing for any queries that filter by UDF in parquet. Are you slowing down
cases where the filter is not selective?
If we want to improve the set of things that we push down, I don't think
specializing for just UDFs in comparison operations is worth it given how much
you are widening the API. Could we just have a single Function filter:
```scala
case class FilterFunction(attribute: String, function: Any => Boolean)
```
or maybe some specialized variants:
```scala
case class IntegerFilter(attribute: String, Int => Boolean)
...
```
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
