Github user cloud-fan commented on the issue:
https://github.com/apache/spark/pull/15807
There is probably a bug of common subexpression elimination: we will
evalute all subexpressions at the very beginning, no matter the results of
subexpressions will be used or not. A counter example:
```
if (isNull(subexpr)) {
...
} eles {
assertNotNull(subexpr)
....
}
```
`assertNotNull(subexpr)` should only be evaluated when `subexpr` is not
null, however, with subexpression elimination, we will always evaluate it.
This may also be bad for performance, if a subexpression is expensive and
won't be evaluted in most cases, but with subexpression elimination, we will
always evaluate it.
cc @davies
---
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]
