Github user HyukjinKwon commented on the issue:
https://github.com/apache/spark/pull/18869
@gatorsmile, I understand the previously discussed point, the documentation
should be correct, tested well and in details. However, if I observed this
correctly, this argument descriptions have not been well filled up for the last
year (since I proposed `Arguments` description) - if I checked correctly, we
have three 3 instances fixed, `replace`, `rlike` and `like`, out of 266
instances.
I mean, practically, it sounds a bit demanding to me. (In this way, for the
perfectness, I think we should also require to add test cases with all the
possible types and check if there is any duplicated test too). I wonder if we
really should add the examples about all the cases we _assume_ users might be
confused of, and describe it that precisely for the current state, given the
information did not exist before.
Of course, it is important and it should be correct but I have been
thinking of incrementally improving this, rather than requesting the perfectly
correct documentation for PRs, practically to roll it better.
What do you think about minimising the test cases and rather trying to
describe it a bit more in an abstract, short and general manner for now? We
could ask additional test cases for additional information in the followup PRs.
For this PR, it looks good enough to me (to be honest, the examples and
test cases look too much) but I believe I need your opinion and I think we
should have a compromised position.
---
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]
