CLN: Add more methods to whitelist (for now)

[jtratner]

No test cases because goal is just to make sure this is not blocked in
0.13. Can do more tweaks later. If I have time I will try to pull
together something to cover what we were discussing in #5480.

[hayd]

"for now" ?

[jtratner]

As in, do it for 0.13 and then figure out about adding test cases and
whatever else later on when we have more time to discuss (since both you
and @gdraps had issues with things not being whitelisted that were supposed
to be).

[hayd]

ok, sweet. I wonder if these are low hanging fruit for vector / cythoning.

[jreback]

@jtratner what's the status of thsi?

[gdraps]

I have a branch with a tentative PR that I can submit later today.

One hitch found while testing is that the whitelist is used for groupby tab
completion but some methods in it are specific to Series or DataFrame
(dtype, corrwith, value_counts). Any thoughts on tailoring the tab
completion based on groupby object type, or should the whitelist be limited
to common methods?
On Nov 27, 2013 10:34 AM, "jreback" [email protected] wrote:

@jtratner https://github.com/jtratner what's the status of thsi?

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/5514#issuecomment-29408564
.

[jtratner]

I'd be inclined to just leave it for both...once you try to use it you'll
learn that it's not available anyways. That way we can keep things more
simple.

[hayd]

The other option is to add it in the Series/DataFrameGroupby class rather than to the whitelist.

[gdraps]

ok, I put a quick fix PR out there: #5604. I used @jtratner's approach of leaving Series/DataFrame-specific methods in the whitelist (dtype, dtypes, boxplot, corrwith, and value_counts). Feel free to use it, or not.

[jreback]

closing in favor of #5604