Github user HyukjinKwon commented on the issue:
https://github.com/apache/spark/pull/15432
I didn't understand. I would like to be very sure on this to avoid mistakes
in my future PRs.
> the current change in this PR also allows `rand` to take an expression
that returns a value of a built-in integer/long data type. Thus, you also need
to test them too, right?
This PR allows `rand` to take an expression but does not affect any other
related language specific functions. Those functions are even already being
tested.
> Not sure whether you are also can check whether `rand` in R and Python
can also correctly handle `NULL`
Assuming from your comment, I can't think of more cases other than the ones
as below (like I said above).
**SparkR**
```r
df <- sql("SELECT RAND(NULL)")
```
**PySpark**
```python
spark.sql("SELECT RAND(NULL)")
df.selectExpr("RAND(NULL)")
```
I see the tests only check if the string is gone well via JVM or not.
IIUC, it is handled in the exact same path with Scala/R/Python within JVM
in this case.
As far as I know, we haven't added/tested such cases in PySpark and SparkR
In history unless it affects each related language specific functions.
Am I missing something?
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
