Remove number of deprecated parameters/functions/classes [fix #6641]

[jsexauer]

Fixes #6641

Line numbers as of commit 70de129:

Deprecated since 0.11 or before

The following are fixed in this PR:

• Deprecate precision in favor of accuracy (Tweak DataFrame formatting to always use same # of digits #395) -- v0.7.0
  .\pandas\core\format.py:{2143}
• Deprecate force_unicode (Unicode : change df.to_string() and friends to always return unicode objects #2224, BUG: series tidy_repr UnicodeDecodeError #2225) -- v0.10.0
  .\pandas\core\format.py:{442, 369}
  .\pandas\core\frame.py:{1387, 1342, 1283}
• Deprecate nanRep in favor of na_rep (nanRep argument deprecation + FutureWarning #275) -- v0.5.0
  .\pandas\core\frame.py:{1285, 1133}
  .\pandas\core\series.py:{910}
• Deprecate unique in HDFStore (ENH: In HDFStore, add select_column method, deprecate unique method #3256) -- v0.11.0
  .\pandas\io\pytables.py:{691}
• Deprecate time_rule favor freq (API deprecations (time_rule -> freq) with moving window functions #1042) -- v0.8.0
  .\pandas\stats\moments.py:{554}
• Deprecate inferTimeRule in favor of infer_freq (Add more robust frequency inference function #391) -- v0.8.0
  .\pandas\tseries\frequencies.py:{247}
• Deprecate DateRange in favor of DatetimeIndex (no issue number, commit 6fe2db5) -- v0.8.0
  .\pandas\core\daterange.py:{24} (will be done in DEPR: Deprecate DateRange #6816)
• Deprecate 'name' in favor of 'symbols' (no issue, commit b921d1a) -- v0.11.0
  .\pandas\io\data.py:{344}
• Deprecate offset in favor of freq (no issue, commit 3136390) -- v0.8.0
  .\pandas\tseries\index.py:{180}

Deprecated since 0.12 or after

Issue #6581 documents the following for future removal. These are not included in this PR

• Deprecate na_last in favor of na_position (EHN/FIX: Add na_last parameter to DataFrame.sort. Fixes GH3917 #5231, BUG: Nested sort with NaN #3917)
  .\pandas\core\series.py:{1770}
• Deprecate rolling_corr_pairwise and expanding_corr_pairwise (Pairwise versions for rolling_cov, ewmcov and expanding_cov #4950)
  .\pandas\stats\moments.py:{301,957}
• Deprecate broadcasting TimeSeries along DataFrame index (Broadcasting bug with datetime.date index? #2304)
  .\pandas\core\frame.py:{2837}
• Deprecate load and save in favor of pickle and to_pickle (CLN deprecate save&load in favour of to_pickle&read_pickle #3787)
  .\pandas\core\common56.py:{2740, 2756}
  .\pandas\core\generic.py:{960, 967}
• Deprecate timeRule and offset in favor of freq (tshift does not work for Panels #4853, Panel tshift 4853 #4864)
  .\pandas\core\datetools.py:{55}
• Deprecate colSpace in favor of col_space (no issue, commit 4a5a677)
  .\pandas\core\frame.py:{1292,1346,1392}
• Deprecate year/month parameters in finance option signatures (Change Finance Options signatures and deprecate year/month parameters #3822, backwards compatible refactoring of io/data options functionality #3817)
  .\pandas\io\data.py:{588,811,848}
• Remove kind from read_excel (CLN: Remove unused and undocumented kind keyword from read_excel #4713)
  .\pandas\io\excel.py:{103}
• Deprecate infer_types to have no effect (REF/BUG/ENH/API: refactor read_html to use TextParser #4770)
  .\pandas\io\html.py:{830}
• Deprecate table keyword in HDFStore (API: remove the table keyword, replaced by fmt='s|t' #4645)
  .\pandas\io\pytables.py:{1124}

[jsexauer]

Wanted to double-check that you want me to remove DateRange as it will be a more significant undertaking than the other ones. In addition to being available at the base of the API, it's used throughout the codebase, so I'll have to refactor all that too. Not a problem to do, but wanted to double-check before doing so.

[jreback]

can certainly leave that for now
but do it a separate pr if u can
it's just really a name for DatetimeIndex

FYI also need to check/update vbenches for any of these deprs

[jsexauer]

Not sure what the vbenches are? Is that what's in the vb_suite directory?

If so, here are the two places with FutureWarning

• https://github.com/pydata/pandas/blob/master/vb_suite/test_perf.py#L601
• https://github.com/pydata/pandas/blob/master/vb_suite/perf_HEAD.py#L92

You don't want me to delete those do you?

[jtratner]

@jsexauer - let's leave out deprecating DateRange for another issue so we can make this a minor change that we can accept. If it's still used within pandas and on the vbenches, it doesn't really feel like DateRange is deprecated.

[jtratner]

i.e. you can just open a separate PR that handles the refactor and deprecation for DateRange.

[jtratner]

this is slightly confusing to me - we deprecated offset as a parameter but not the actual offset property on the object?

[jsexauer]

Correct; there is no notation I can find that says the offset property of the pandas Timestamp object is deperated. However, freq is an alias of offset:

https://github.com/pydata/pandas/blob/master/pandas/tslib.pyx#L258

[jsexauer]

Ok I will do the DateRange one next. Just made a small change I saw in updating the docs (shouldn't affect tests). I think this one is ready for closing; let me know if more work is needed.

[jreback]

can you edit #6581 to update the >= 0.12 (that we are going to deprecate in the future)?

thanks

[jreback]

anyone have a problem with any of this, these are all removed deprecations that were in effect <= 0.11 (so a while ago). That were announced to be removed in a future version.

@jsexauer @jtratner @jorisvandenbossche @TomAugspurger @cpcloud @hayd @wesm

[cpcloud]

Looks good, but I don't see the html.py removal listed in the task list.

[jsexauer]

@cpcloud It is the second to last item in the "Deprecated since 0.12" list. I see that the warning message says to remove in 0.14, but warning was given only in 0.13. Let me know if you want me to remove anyways.

[cpcloud]

oh ok ... you're right ... sorry i haven't looked at html stuff in a while, carry on

[jreback]

yep....only doing < 0.12 announced deprecations (which their are a lot!)

[wesm]

I'm pretty wary of removing load and save since they're used in The Book. I'm not sure if any of these other deprecation will affect the book but should try to avoid if possible, I guess.

[jreback]

save/load not on the table as those where from. 0.12

@jsexauer in the list can u out the version that the depr happened (I know I have the commit but for reference) thanks

[jsexauer]

@wesm I don't own a copy of The Book unfortunately, but assuming it was written for v0.9 (given publish in Oct 2012) the only two of these that might affect it are

• the removal of the unique keyword from HDFStore.select_column
• deprecation of name in favor of symobls in pandas.io.data.get_data_yahoo and pandas.io.data.get_data_google

Do you know if either of these were used?

@jreback Modified inital PR comment to include versions of first deprecation. Also moved all deprecations not pulled out in this PR into #6581.

[hayd]

Neither of those are used in wes' book.

[jreback]

yep...i think this is all fine.

[jreback]

thanks @jsexauer

[jreback]

@jsexauer ok..this is merged, but can you do a follow up fix the issues in v0.14.0 (and prob in release.rst), so that they are the original issues numbers (and not the deprecation issue).

http://pandas-docs.github.io/pandas-docs-travis/whatsnew.html#whatsnew-0140-prior-deprecations

e.g. if you are looking thru that page you want to be able to click thru and see what the original issue was.

[jsexauer]

What should I do for the ones that that didn't have issues?

[jreback]

if u have a commit reference then u can use that (though u may have to do an inline reference)
if u can get it to work great if leave the existing referemce

[jsexauer]

Have an example you could point me to?

[jreback]

oh...looking in cookbook.rst, its just another external link. (The issues render differently/automatically, via the issue number), but a commit is just a reference like any other HTML link.