Github user mgaido91 commented on the issue:
https://github.com/apache/spark/pull/19860
@viirya sorry, I don't understand your question.
In Coalesce, we need to find the first non-null element. As soon as we find
one, we don't need to evaluate anything else. Previously, the code generated by
coalesce would have been:
```
methodName_1();
methodName_2();
...
methodName_X();
```
and in each method we were using `${ev.isNull}` to avoid the computation of
the unnecessary expressions, after the first non-null condition was met.
In this case, even though we are doing nothing inside these function we are
still calling all them and this is not cheap, as pointed out by @gatorsmile
here: https://github.com/apache/spark/pull/19752#discussion_r153081547.
Thus, in the new generated code, we avoid calling the methods when it is
not necessary, since the generated code is:
```
do {
methodName_1();
if (!isNull_1234) {
continue;
}
...
} while (false);
```
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
