Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Fixups for categoricals #8153

Merged
merged 7 commits into from
Sep 25, 2014
Merged

Fixups for categoricals #8153

merged 7 commits into from
Sep 25, 2014

Conversation

jankatins
Copy link
Contributor

This PR will collect the various fixups for Categorical.

See #8078 for the ToDos.

Closes: #8077 -> API/FIX: wrap result cut/qut in Series and make it ordered
Closes: #8074 -> Discussion
Closes: #8078 -> Todos

This has a (passing) unittest for #8138 but no code changes. Not sure what was the change which fixed the bug there...

@jorisvandenbossche
Copy link
Member

@JanSchulz Thanks for working on this! But, maybe it is simpler that you just put different PRs instead of one big. Then we can also more quickly merge if one part is ready, not waiting on other issues. There is now a clear to do list, and some of the issues are not necessarily dependent on each other. Eg I would leave this PR just as a PR to fix cut/qcut

jreback
jreback reviewed Sep 1, 2014
View reviewed changes
pandas/tools/tile.py
return _bins_to_cuts(x, bins, right=right, labels=labels,
retbins=retbins, precision=precision,
include_lowest=include_lowest)
if isinstance(x, Series):
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

you should do the conversion to series after bins_to_cuts is done

eg

res = _bins_to_cuts(...)
if isinstance(x, Series):
res = Series(res)
return res

@jankatins
Copy link
Contributor Author

I think that's fine with this change but the rest will become dependend on each other. I will add a new PR for the rest but would like to do the rest in one PR (I will do one commit per change, hope that helps)

@jorisvandenbossche
Copy link
Member

OK, do as you seem fit, no problem! (but this can indeed go in as a seperate PR)

@jankatins jankatins force-pushed the categorical_fixups_2 branch from 0b631ce to 05623d7 Compare September 1, 2014 12:55
jorisvandenbossche
jorisvandenbossche reviewed Sep 1, 2014
View reviewed changes
pandas/tools/tests/test_tile.py
exp = Series(Categorical.from_codes([0,0,0,1,1,1,2,2,2],
["(-0.008, 2.667]", "(2.667, 5.333]", "(5.333, 8]"],
ordered=True))
tm.assert_series_equal(res, exp)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

assert_series_equal does not check for the orderedness of the Categorical at the moment, so the test will not check that part of the issue:

In [125]: c1 = pd.Series(pd.Categorical(list('abcdaba'), ordered=True))

In [127]: c2 = pd.Series(pd.Categorical(list('abcdaba'), ordered=False))

In [129]: pd.util.testing.assert_series_equal(c1, c2)

But maybe that is for a seperate issue if this should test that or not.

@jankatins jankatins force-pushed the categorical_fixups_2 branch from 05623d7 to f20553b Compare September 1, 2014 13:55
@jorisvandenbossche jorisvandenbossche mentioned this pull request Sep 1, 2014
Closed
@jankatins jankatins force-pushed the categorical_fixups_2 branch from f20553b to 6919aea Compare September 1, 2014 21:49
@jankatins jankatins mentioned this pull request Sep 1, 2014
Closed
@jankatins jankatins force-pushed the categorical_fixups_2 branch 2 times, most recently from 4f08b65 to a245833 Compare September 2, 2014 12:02
@jorisvandenbossche
Copy link
Member

@JanSchulz some feedback on the ENH: let Categorical.reorder_levels also remove levels commit. I think that is a nice feature, but I don't know I would use that as default. In a sense that, if I only want to 'reorder' the levels (the 'core' functionality of the function), it would be nice that the function warns me when I eg forgot a level. Now it will silently drop those.

So just spawning ideas:

  • have a strict keyword (or other name), that if True -> only same length is allowed, and if False -> you can also add/remove items in one go. If this would be True by default, then such removing of levels by accident cannot happen, as the user should set this to False to specifically say: 'Yes, I want to add/remove levels and I am myself responsible to check that I provided the correct items'. The error message when this kwarg is True could also point to set it to False.
  • have seperate functions like add_levels, remove_levels and keep reorder_levels as is. Or maybe these would be useful as such, regardless of this discussion. Because, if you just want to add or remove a level (without reordering in the same time), this could be simpler as using reorder_levels where you also have to specify the others.

@jankatins
Copy link
Contributor Author

I like the strict idea. Will amend that commit.

@jorisvandenbossche
Copy link
Member

for the name strict, @jreback do you know of other places in pandas where a similar functionality is used? (to have consistent naming in that case)

I also think my second bullet (the add_levels/remove_levels function) could also be useful irregardless of the behaviour of reorder_levels

@jreback
Copy link
Contributor

jreback commented Sep 3, 2014

strict is ok, not really sure its used elsewhere. I think you'd rather have the nomeclature of set/reorder indexes though?

e.g.

reorder_levels (change the ordering, must have same number of levels)
set_levels (can say any and can actually change the ordering / no check on number of levels)
reset_levels (I guess this is remove)?

jreback
jreback reviewed Sep 3, 2014
View reviewed changes
pandas/core/categorical.py
warn("Values and Levels have different dtypes. Did you mean to use\n"
"'Categorical.from_codes(codes, levels)'?", RuntimeWarning)

if com.is_integer_dtype(values) and (codes == -1).all():
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why is this on the fastpath? the point of the fastpath is that almost no checking is done (and its completely internal), so it is truly the fast path

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's not, it just looks like this in the diff (fastpath returns in line 206)

@jreback jreback added this to the 0.15.0 milestone Sep 3, 2014
@jankatins
Copy link
Contributor Author

@jreback @jorisvandenbossche re Strict: There are actually 4 types of actions you can make:

  1. rename levels (currently done by assigning new things to levels)
  2. removal of levels (can be done by reorder_levels on any place in the levels and by assignments to levels at the last places)
  3. additions of new levels (can also be done by both: at any place by reorder_levels and after the last place by assignments to levels
  4. reorder current levels (by reorder_levels)

So, the "strict" idea is actually very good, but if there should be a operation per "action", then rename, remove and reorder can be done without "sideeffects", but additions not: should that take a list of only the new levels (and put that at the last/first place in the list?) or should that take a list of all levels and it's then a kind of reorder_levels, because it can insert at any place?

One "not nice" thing is, that each of this sideeffect free operation would actually take quite a lot of checks which make it saver but are actually unnecessary if we would allow them...

@jreback
Copy link
Contributor

jreback commented Sep 9, 2014

@JanSchulz how's this coming?

@jankatins
Copy link
Contributor Author

I will work on this over the weekend. Sorry for the delay...

@jankatins
Copy link
Contributor Author

I will take the "strict" path for reorder. I really don't like the idea of adding a method per action (remove, add, reorder, rename) which in the end would all be some kind of "prevent side effects"-checks and then call some kind of method which would be like the current reorder_levels(). So the current end result is:

  • reorder_levels( ... strict=True) -> can only reorder levels, will check that the set of new levels are the same as the set of current levels
  • reorder_levels( ... strict=False) -> can reorder levels, including adding new ones (will be unused) or remove some (will become -1 = NaN)
  • assigning to levels -> renames levels, can be used to remove/add levels from the end of the list.

I'm actually considering to make assigning to levels stricter, so that it only can rename, without the possibility to add/remove levels.

Anny comments?

@jorisvandenbossche
Copy link
Member

I have the feeling that the add/remove/reorder_levels needs some more discussion. So maybe remove that one commit from this PR? (the one to last commit 'ENH: let Categorical.reorder_levels also remove levels … ')
So we can review/try to get the rest of this PR merged as soon as possible, and the we can discuss it further somewhere else (I can create a new issue).

@jankatins
Copy link
Contributor Author

I can put that at the end, but I would really like to get the rename (levels -> categories) done in this PR and this will touch everything...

@jorisvandenbossche
Copy link
Member

OK, the possible addition of other methods like add/remove_levels is maybe complementary to this and this does not exclude that, so maybe not a problem. (although I still think adding seperate methods can be useful, but lets keep that for somewhere else)

+1 on assigning to levels be strict (so check for length, only rename)

jreback
jreback reviewed Sep 23, 2014
View reviewed changes
pandas/core/categorical.py
>>> s.cat.reorder_levels(list('cab'))
>>> s.cat.categories
>>> s.cat.categories = list('abc')
>>> s.cat.reorder_categories(list('cab'))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

maybe add even more examples here (and mention that these are INPLACE). - can be another pr if you want

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

only the assignment to categories is inplace.

Willl add some examples and a note about inplace/not behaviour.

@jreback
Copy link
Contributor

jreback commented Sep 23, 2014

@JanSchulz pretty good!

pls just address (relatively minor) doc changes, and then can get this in. lmk as soon as you are ready. don't 'add' anything else. let's just get in and close the open issues. can do (small followups) if necessary.

@jankatins
Copy link
Contributor Author

@jreback ok, will do. Thanks for the review!

@jreback
Copy link
Contributor

jreback commented Sep 23, 2014

btw, in the very first doc section (and v0.15.0). I think if you could put a 'real' example of why someone would use a categorical, e.g. maybe sorting by the categorical in a particular order to show data a certain way - sort of the motivational example (no need to really explain it, just sort of show it so people get the ahah of why one would use a categorical)

@jankatins
API/FIX: wrap result cut/qut in Series and make it ordered
ca2e608 
As per the discussion in pandas-dev#8077, the result of cut/qcut should be
wrapped into a Series, now that this is possible.

Also change the returned Categorical to "ordered=True", this was
overlooked when this was introduced in Categorical.

Closes: pandas-dev#8077
@jankatins
Copy link
Contributor Author

I can't get my pandas 3.4 environment working, so lets see what travis puts out.

@jankatins
Copy link
Contributor Author

I added all the changes to the last commit.

@jreback
Copy link
Contributor

jreback commented Sep 23, 2014

ok, once (hopefully) green will merge.

@jankatins
Copy link
Contributor Author

It's green, but I'm not sure why, I didn't change something in the code, only wanted to make the tests more clear.

@jreback
Copy link
Contributor

jreback commented Sep 24, 2014

ok, pls squash down a bit (don't need to go to 1 but just narrow down for clarity) pls.

@jreback jreback mentioned this pull request Sep 24, 2014
Closed
@jankatins
FIX: Implement small Categorical fixups
206fb97 
This is a squashed commit which adresses several small issues
identified in the latest code review.

* As per dicussion in pandas-dev#8074,
  Categorical.unique() should return an array and not a Index object
  to make is consistant with other unique() implementations.

* Mark a few more categorical methods as internal

* Change reorder_levels and drop_unused_levels to inplace=False. As per
  the discussion in pandas-dev#8074, change the default of both methods
  to return a copy and only do the change inplace if told otherwise.

  Discusion in
  pandas-dev#8074 (comment)

* Make str(cat) more array like. As Categorical is more like
  np.ndarray and less similar to pd.Series, change the string
  representation of Categorical to be more like np.ndarray.

  The new str(cat) will only show at max 10 values.

  Also remove the "Categorical(" in front of an empty categorical
  -> either that's put in front of every str(cta) or none. It also missed
  the final ")".

* Test more nan handling in Categorical

* Add a testcase for groupby with two columns and unused categories

  During the devleopment of pandas new Categorical support, groupby
  with two columns (one worked) didn't include empty categories which
  omits rows with NaN. Add a testcase so this is checked.

  No codechanges nessesary, this bug was fixed at some other places
  before. See pandas-dev#8138 for more details.
@jankatins
API: remove compat keyword in Categorical constructor
8fd654c 
Before the new Categorical work, the default two argument constructor
was expecting "codes and levels" but this was changed to
"values and levels" and a 'compat' kwarg was added, which could
be used to switch to the old style constructor useage.

It was decided that we switch to the new style constructor useage as
the new default (compat=False) and implement a 'from_codes(...)'
constructor.

As the compat codepath is now never triggered and code should be
changed to use the new 'from_codes()' constructor, remove the old
codepath.

Add some warnings if we are pretty sure that the old style
constructor is meant, but unfortunately we don't catch all cases.
@jankatins
API: Implement Series.cat.codes and rename the proxy class
b74e0a3 
* Series should also expose Series.cat.codes as a Series of codes.
  See discussion in
  pandas-dev#8074 (comment)

* Rename CategoricalProperties CategoricalAccessor. It has more
  than properties, so "accessor" is more appropiate.
@jankatins
API: rename Categorical.levels to .categories
9d86247 
The name 'levels' was already used by a much different concept in
MultiIndex and this was too confusing, so change the name to
'categories'. Add deprecation warning if the old name is used
(in constructor or the public 'levels' attribute). The old name
'levels' is not anymore exposed unter Series.cat as it was never
part of a stable release.

See the discussion in pandas-dev#8074

This rename was done by search&replace in categorical.py and the
corresponding tests_categorical.py, implementing the deprecation
accessor (with a temporary 'raise Exception') and then run the the
unittests and change code until all tests pass.
@jankatins
ENH: Introduce methods for each change to Categorical.categories
4c45d7f 
There are multiple "actions" which can be done to categories:

- add one or more categories
- remove one or more categories
- rename the categories
- reorder the categories
- remove unused categories

Each of these actions now gets it's own method together with a special
`set_categories`, which can do more than one action in one call, which is
needed to save computation time.

See pandas-dev#8153 (comment) and the
following comments.
@jankatins
DOC: update to reflect changes in Categorical
3f282da 
* rename of levels -> categories
* don't use the `Categorical` constructor if thats not needed
* introduce the new methods to manipulate categories
@jankatins
Copy link
Contributor Author

Ok, did it, but IMO the commits are now not as orderly as before: before one commit = one logical change, now they are mostly mixups of multiple changes... :-(

@jreback
Copy link
Contributor

jreback commented Sep 25, 2014

ok this looks fine

jreback added a commit that referenced this pull request Sep 25, 2014
@jreback
Merge pull request #8153 from JanSchulz/categorical_fixups_2
0ffef1d 
Fixups for categoricals
@jreback jreback merged commit 0ffef1d into pandas-dev:master Sep 25, 2014
@jreback
Copy link
Contributor

jreback commented Sep 25, 2014

@JanSchulz thanks!

pls review docs after they r built

@jreback jreback mentioned this pull request Sep 25, 2014
Closed
17 tasks
bboerner pushed a commit to bboerner/pandas that referenced this pull request Oct 23, 2014
Squashed commit of the following:
d85cdec 
commit 7012d71c61cc4fc8e47e231659a5ca01629ce1ae
Merge: 8e0155c 6fb8f14
Author: Joris Van den Bossche <[email protected]>
Date:   Thu Oct 23 08:57:57 2014 +0200

    Merge pull request #8606 from jorisvandenbossche/doc-whatsnew-reorg

    DOC: reorg whatsnew files to subfolder

commit 8e0155cb9be4805cdd650479ea03a040d7c2b5c9
Merge: bd01395 00965aa
Author: jreback <[email protected]>
Date:   Wed Oct 22 16:47:38 2014 -0400

    Merge pull request #8604 from aisipos/fixplotdoctypo

    Fix typo in visualization.rst doc

commit 6fb8f14ab4517fe5e5b1bdb5a4478af16ddcfab2
Author: Joris Van den Bossche <[email protected]>
Date:   Wed Oct 22 08:55:47 2014 +0200

    DOC: reorg whatsnew files to subfolder

commit bd0139505c61f52dd75d12f5d04982ebe5e383ac
Merge: 8747633 0173962
Author: Joris Van den Bossche <[email protected]>
Date:   Wed Oct 22 22:43:54 2014 +0200

    Merge pull request #8583 from jorisvandenbossche/doc-upload

    DOC/REL: adapt make.py file to upload with other user than pandas

commit 00965aa2b66f01ba0d39425f94a1a8319e6d8133
Author: Anton I. Sipos <[email protected]>
Date:   Wed Oct 22 12:25:59 2014 -0700

    Fix typo in visualization.rst doc

commit 8747633f2ac7ec3c38eab9debb4f4ab0da13259c
Merge: 3c1185f df24187
Author: jreback <[email protected]>
Date:   Wed Oct 22 06:54:40 2014 -0400

    Merge pull request #8599 from shoyer/qualify-mem-usage

    Use '+' to qualify memory usage in df.info() if appropriate

commit df2418774861111238bdca1edfb8908fe9db3da9
Author: Stephan Hoyer <[email protected]>
Date:   Wed Oct 22 00:30:54 2014 -0700

    Use '+' to qualify memory usage in df.info() if appropriate

    This should resolve @kay1793's complaint in #8578.

commit 3c1185fd03306a65101657c32b58e4f59c3d1376
Merge: 403f38d 2dd86e8
Author: jreback <[email protected]>
Date:   Tue Oct 21 19:20:26 2014 -0400

    Merge pull request #8598 from artemyk/io_doc_fix

    Function pointer misspecified in io documentaion

commit 2dd86e812abf691ac433cb6bf03f4bd3419a6f5f
Author: Artemy Kolchinsky <[email protected]>
Date:   Tue Oct 21 17:38:14 2014 -0400

    Documentation fix

commit 403f38da35ab04d3997f61db3c98134fe02910d3
Author: pallav-fdsi <[email protected]>
Date:   Sun Oct 19 20:49:21 2014 -0400

    BUG: invalid import checking in gbq (GH8587)

commit a2c36c1354ae8d37356fd3c78bb2406546e7d371
Merge: 8336e36 902179d
Author: jreback <[email protected]>
Date:   Mon Oct 20 07:03:40 2014 -0400

    Merge pull request #8577 from bashtage/stata-columns

    ENH: Allow columns selection in read_stata

commit 902179dc59c24e0f0820ffb25fb1ecf7e04c401d
Author: Kevin Sheppard <[email protected]>
Date:   Fri Oct 17 20:16:47 2014 -0400

    ENH: Allow columns selection in read_stata

    Allow columns to be selected when importing Stata data files.

    Closes #7935

commit 8336e362f930d843c7f80061e0f8f99ed5d3267f
Merge: 778cfe4 cf599d9
Author: jreback <[email protected]>
Date:   Sun Oct 19 20:58:25 2014 -0400

    Merge pull request #8575 from immerrr/optimize-makestringindex

    TST/PERF: optimize tm.makeStringIndex

commit 0173962e9696413f0294b1f13232899330e8bc2b
Author: Joris Van den Bossche <[email protected]>
Date:   Sun Oct 19 22:18:14 2014 +0200

    DOC/REL: adapt make.py file to upload with other user than pandas

commit 778cfe43a4cef52442e00fe505290079a717f515
Author: Jeff Reback <[email protected]>
Date:   Sat Oct 18 18:58:29 2014 -0400

    DOC: start 0.15.1 cycle for docs

commit 017adeaa4b70d63fe6c788db457dc9d31562f4d6
Author: Jeff Reback <[email protected]>
Date:   Sat Oct 18 18:52:16 2014 -0400

    RLS: 0.15.0 final

commit de4b0bb202210c0b74a6a3b48e05ef39505d0670
Merge: 316f9fa 4e40a64
Author: Joris Van den Bossche <[email protected]>
Date:   Sun Oct 19 00:42:16 2014 +0200

    Merge pull request #8579 from jorisvandenbossche/doc-fixes015

    DOC: fix example of GH8512 name clash with following examples

commit 4e40a64330c32d9d994f82254252bfcefc6d67b8
Author: Joris Van den Bossche <[email protected]>
Date:   Sat Oct 18 23:53:33 2014 +0200

    DOC: fix example of GH8512 name clash with following examples

commit 316f9facc040829f4633cc76683824ffaf49221c
Author: jreback <[email protected]>
Date:   Sat Oct 18 08:59:40 2014 -0400

    DOC: updates release and v0.15.0.txt for release

commit 3788c65735445eb2142d83fdd7e7633c7f6a7b43
Author: Phillip Cloud <[email protected]>
Date:   Fri Oct 17 20:01:18 2014 -0400

    BUG: dtype query with negation bug (GH8568)

commit cf599d9d6fd5734da313e50fc47da45a464b2b7e
Author: immerrr <[email protected]>
Date:   Thu Oct 16 14:36:40 2014 +0400

    TST/PERF: optimize tm.makeStringIndex

    * add tm.rands_array(nchars, size) function to generate string arrays

    * add tm.randu_array(nchars, size) function to generate unicode arrays

    * replace [rands(N) for _ in range(X)] idioms:
      - with makeStringIndex in benchmarks to maintain backward compatibility
      - with rands_array in tests to maintain 1-to-1 type correspondence

commit 2baaefe406aad2b4e809f9e6605f5f05f5dad90d
Merge: ec31259 396e7f3
Author: jreback <[email protected]>
Date:   Fri Oct 17 03:22:32 2014 -0400

    Merge pull request #8564 from bashtage/stata-datatype-conversion

    ENH: Add type conversion to read_stata and StataReader

commit 396e7f3b8a3a8d9910fdc193685ae3bf75b79479
Author: Kevin Sheppard <[email protected]>
Date:   Sun Oct 12 16:46:21 2014 -0400

    ENH: Add type conversion to read_stata and StataReader

    Added upcasting of numeric types which will default numeric data to either
    int64 or float64.  This option has been enabled by default.

    closes #8527

commit ec31259e10396056317e9964ac48d128472ca76f
Merge: 09e0ba7 62913bc
Author: jreback <[email protected]>
Date:   Thu Oct 16 21:19:54 2014 -0400

    Merge pull request #8570 from jreback/equals

    PERF: improve perf of array_equivalent_object (GH8512)

commit 62913bc1a6c2b0a6fd492609dea18a4797500a38
Author: jreback <[email protected]>
Date:   Thu Oct 16 20:36:36 2014 -0400

    PERF: improve perf of array_equivalent_object (GH8512)

commit 09e0ba73c1ff702adcb1349fca6f839b44857fb7
Merge: 5058224 54f047a
Author: jreback <[email protected]>
Date:   Thu Oct 16 12:12:20 2014 -0400

    Merge pull request #8566 from shoyer/fix-default-cmap

    Switch default colormap for scatterplot to 'Greys'

commit 50582247269937fa1fc2e861c9b6424cf546f82e
Merge: 5cd1bed f97c1d9
Author: jreback <[email protected]>
Date:   Thu Oct 16 08:49:02 2014 -0400

    Merge pull request #8567 from seth-p/ignore_na_warning

    DOC: ignore_na warning

commit 54f047a564433baedc9da0fea3e78e11985afaf4
Author: Stephan Hoyer <[email protected]>
Date:   Wed Oct 15 17:50:52 2014 -0700

    Switch default colormap for scatterplot to 'Greys'

    In #7780, I added colormap support to scatter, but choose to override the
    default rainbow colormap (7780#issuecomment-49533995).

    I'm now regreting choosing 'RdBu' as the default colormap. I think a simple
    black-white colormap is a more conservative and better default choice for
    coloring scatter plots -- binary colormaps really only make sense if the data
    is signed. 'Greys' is also the default colormap set by the seaborn package,
    and @mwaskom has done far more thinking about colormaps than I have.

    If possible, I would love to get this in before 0.15 is released, so we don't
    have to worry about changing how anyone's plots look.

    CC @sinhrks

commit f97c1d95d898df74e8b137222190c8e66aca0000
Author: seth-p <[email protected]>
Date:   Wed Oct 15 21:55:39 2014 -0400

    DOC: ignore_na warning

commit 5cd1bedd56f5ba038cf29c7eaaee69f82c422267
Merge: e0d9b1a acb9b33
Author: jreback <[email protected]>
Date:   Wed Oct 15 14:21:23 2014 -0400

    Merge pull request #8549 from mtkni/master

    PERF: Slowness in multi-level indexes with datetime levels (part 2)

commit e0d9b1a7807c41374b0212c3b09d5b73c1fecec6
Author: John W. O'Brien <[email protected]>
Date:   Sun Sep 21 21:43:56 2014 -0400

    BUG: _Openpyxl2Writer mishandles merged cell style

    TST: Cover merged cells case in openpyxl2 engine

commit 4737f6ce83d55902b2af5df301a9b86a48ed0c64
Merge: 51bd2c9 0e6a858
Author: jreback <[email protected]>
Date:   Wed Oct 15 13:35:37 2014 -0400

    Merge pull request #8561 from JanSchulz/r_levels

    DOC/FIX: In R it's still levels...

commit 0e6a858390eb33eac89589cd7fa406d0b2ea5c89
Author: Jan Schulz <[email protected]>
Date:   Wed Oct 15 18:25:17 2014 +0200

    DOC/FIX: In R it's still levels...

commit 51bd2c901ab437a26addad37ebacdbaecd840b6e
Author: Joris Van den Bossche <[email protected]>
Date:   Wed Oct 15 14:08:21 2014 +0200

    DOC: formatting issue in scatter_matrix section

commit 5141ef5f66f3b0b8453699d0fe089214e5ee684c
Merge: e9f55a3 857042f
Author: jreback <[email protected]>
Date:   Tue Oct 14 21:24:00 2014 -0400

    Merge pull request #8557 from jreback/dt_perf

    PERF: fixup datetime_index repr perf (GH8556)

commit 857042ff21323bdc56ca6e8e66a59248fe8f3116
Author: jreback <[email protected]>
Date:   Tue Oct 14 20:39:57 2014 -0400

    PERF: fixup datetime_index repr perf (GH8556)

commit acb9b3390c8c7df8275b7a7463f34bc640a52fba
Author: Mike Kelly <[email protected]>
Date:   Mon Oct 13 10:18:15 2014 -0400

    PERF: Slowness in multi-level indexes with datetime levels

    Special case handling of sliced multi indexes, where there the length of the level values may exceed the length of the index.
    Corrected nan handling issue introduced by the original change. (+1 squashed commit)
    Squashed commits:
    [0dae170] PERF: Slowness in multi-level indexes with datetime levels

    Special case handling of sliced multi indexes, where there the length of the level values may exceed the length of the index.

commit e9f55a3aa0932a1b0a496e2c0bd02443e8d3cd80
Merge: 69e0f91 c71c536
Author: jreback <[email protected]>
Date:   Mon Oct 13 10:20:11 2014 -0400

    Merge pull request #8507 from westurner/patch-3

    DOC: Update ecosystem.rst: Add IDE, IPython, qgrid, Spyder, API, quandl, Out-of-core, Blaze

commit c71c5360f0476ee0a71ae1f41b02b32977975c2f
Author: Wes Turner <[email protected]>
Date:   Wed Oct 8 05:57:58 2014 -0500

    DOC: Update ecosystem.rst: Add IDE, IPython, qgrid, Spyder, API, quandl, Out-of-core, Blaze

    DOC: Update ecosystem.rst: Add IDE, IPython, API
    DOC: Update ecosystem.rst: Add GUI > Spyder
    DOC: Update ecosystem.rst: GUI -> IDE, IPython > qgrid, API
    DOC: Update ecosystem.rst: -> IDEs
    DOC: Update ecosystem.rst: depluralize
    DOC: Update ecosystem.rst: Add IPython [Notebook] description; alphabetize
    DOC: Update ecosystem.rst: add Blaze, alphabetize

commit 69e0f91239c7943674c438e681b11161faa831e3
Author: jreback <[email protected]>
Date:   Mon Oct 13 07:55:04 2014 -0400

    DOC: v0.15.0 corrections

commit 67aa5c80b45d3f356de7e574bef2f1359891edbf
Merge: f5fe192 2c4dc5c
Author: jreback <[email protected]>
Date:   Mon Oct 13 07:36:39 2014 -0400

    Merge pull request #8544 from mtkni/master

    PERF: Performance improvement for MultiIndexes with a DatetimeIndex levels

commit f5fe1927c936b0572df702754bf1f8324e62ef57
Merge: 2e8f236 42cde88
Author: jreback <[email protected]>
Date:   Sun Oct 12 22:19:33 2014 -0400

    Merge pull request #8535 from jnmclarty/cookbook_inline_examples

    DOC: Clean up Sphinx Issues, fix tabs, in Cookbook

commit 42cde8817ae0608b1771221c86b0c9df0039bac3
Author: jnmclarty <[email protected]>
Date:   Fri Oct 10 22:11:23 2014 -0400

    DOC: Clean Cookbook Issues: Tabs, expanded_apply, axis aliasing, etc.

commit 2c4dc5c74c2485a28450804d078ca872e1494868
Author: Mike Kelly <[email protected]>
Date:   Sat Oct 11 17:14:48 2014 -0400

    PERF: Performance improvement for MultiIndexes with a DatetimeIndex level

commit 2e8f236388af12b207254bd7143147dc545aa8b9
Merge: e560f9b 1054ed4
Author: Tom Augspurger <[email protected]>
Date:   Sat Oct 11 14:27:25 2014 -0500

    Merge pull request #8536 from TomAugspurger/plot-label-none

    BUG: Dont add None to plot legend.

commit 1054ed4192cb3014b4dd7e5078b823049d9b4e36
Author: TomAugspurger <[email protected]>
Date:   Sat Oct 11 07:29:55 2014 -0500

    BUG: Dont add None to plot legend.

commit e560f9b8dd5732edf4d93dbf938abc40fbc87a83
Merge: ef6ce1a 58e8cdd
Author: jreback <[email protected]>
Date:   Fri Oct 10 19:09:30 2014 -0400

    Merge pull request #8288 from jnmclarty/cookbook_inline_examples

    Cookbook Inline Examples 2014Q3

commit 58e8cdd179a88d873f223cff3849d3580d326ca8
Author: jnmclarty <[email protected]>
Date:   Sun Sep 14 10:34:54 2014 -0400

    DOC: Add 29+ in-line examples, change headings & intro to cookbook GH6918

commit ef6ce1aa3099c5bc78c2e8c6b307356fe67a0497
Merge: 56dbb8c 3493550
Author: Joris Van den Bossche <[email protected]>
Date:   Sat Oct 11 00:32:42 2014 +0200

    Merge pull request #8534 from shoyer/add-xray-to-ecosystem-docs

    Add xray to ecosystem docs

commit 349355049e186efecda74a0b48a530948d71b798
Author: Stephan Hoyer <[email protected]>
Date:   Thu Oct 9 23:23:32 2014 -0700

    Add xray to ecosystem docs

    Fixes #8516

commit 56dbb8cc0cf810bdf9af472f16f6f14c5393e8ee
Merge: ce79c80 b68017e
Author: jreback <[email protected]>
Date:   Fri Oct 10 08:50:13 2014 -0400

    Merge pull request #8519 from immerrr/fix-categoricalblock-pickling

    BUG: fix CategoricalBlock pickling

commit ce79c80578708c32a161b9e109c8804ebbf5cea5
Author: thatneat <[email protected]>
Date:   Tue Oct 7 09:40:48 2014 -0700

    API: Raise more specific exception on concat([None])

commit b68017eb59ea46e6e6716ac2490c9530eb3510d9
Author: immerrr <[email protected]>
Date:   Thu Oct 9 15:05:08 2014 +0400

    BUG: fix CategoricalBlock unpickling

    TST: add categorical frame and series to generate_legacy_pickles

    TST: generate legacy pickle including frames/series with categoricals
    for 0.15.0

    COMPAT: ensure pre-0.15.0 pickles of Categorical are read

    TST: move assert_categorical_equal to tm.TestCase

commit 32c5016de21d4f0c31f1c7eee7b98a43231b9fda
Merge: 520170d 8a662c6
Author: jreback <[email protected]>
Date:   Thu Oct 9 21:17:56 2014 -0400

    Merge pull request #8512 from unutbu/equal-warning

    BUG: Suppress FutureWarning when comparing object arrays

commit 520170d5ab5a3ab488b75688f5c6295104542cbc
Author: jreback <[email protected]>
Date:   Thu Oct 9 19:59:14 2014 -0400

    CI: appveyor fixes

commit 8a662c6ea4700420c4af0d6b3992f57d172ee9e4
Author: unutbu <[email protected]>
Date:   Thu Oct 9 19:44:37 2014 -0400

    BUG: Avoid FutureWarning by implementing equality check in Cython.
    BUG: array_equivalent expects NumPy arrays as input.
    DOC: Fix typo and show that the index order matters

commit 2d4c8e5ccc84b94ae3d2f01ecbad96d8b31d7493
Author: jreback <[email protected]>
Date:   Thu Oct 9 15:10:30 2014 -0400

    CI: incoporate appveyor build folder variable

commit 6cc875ce4e5a6b1e3afd720a9c0f8f186154f3da
Merge: 89e6092 574facc
Author: jreback <[email protected]>
Date:   Thu Oct 9 14:32:16 2014 -0400

    Merge pull request #8523 from jreback/regr

    BUG/REGR: bool-like Indexes not properly coercing to object (GH8522)

commit 574facc76c898377a11956f1a7051de34965efbe
Author: jreback <[email protected]>
Date:   Thu Oct 9 13:10:45 2014 -0400

    BUG/REGR: bool-like Indexes not properly coercing to object (GH8522)

commit 89e60925b173109ea5c45ea7786b53b4c1b128fe
Merge: a05b8ed 9e0122a
Author: jreback <[email protected]>
Date:   Thu Oct 9 09:36:59 2014 -0400

    Merge pull request #8520 from jreback/indexing

    BUG: Bug in inplace operations with column sub-selections on the lhs (GH8511)

commit 9e0122a31807847dad8ce06bef42899b5f6454dc
Author: jreback <[email protected]>
Date:   Wed Oct 8 19:05:15 2014 -0400

    BUG: Bug in inplace operations with column sub-selections on the lhs (GH8511)

commit a05b8ed17295f952f58a093a029c7b97d3e6224d
Merge: aaba591 d2bc2fe
Author: Joris Van den Bossche <[email protected]>
Date:   Wed Oct 8 22:14:52 2014 +0200

    Merge pull request #8510 from jorisvandenbossche/doc-fixes

    DOC: fix example sql chunksize

commit d2bc2fec54bc97c57f28f8d75a39d5ef9c35fb82
Author: Joris Van den Bossche <[email protected]>
Date:   Wed Oct 8 17:15:55 2014 +0200

    DOC: fix example sql chunksize

commit aaba5915bf89a0d7eb0ab3434354d9448b1297ff
Author: jreback <[email protected]>
Date:   Wed Oct 8 10:07:08 2014 -0400

    DOC: merging.rst changes

commit 215569ac8fdb4428a544054635bc94fc296293f3
Merge: e1083ee d9ee560
Author: jreback <[email protected]>
Date:   Tue Oct 7 17:52:42 2014 -0400

    Merge pull request #8500 from jreback/appv

    WIP/CI: add appveyor support

commit d9ee560dd4cb83d3628fdfd7b6f9d420628aab42
Author: jreback <[email protected]>
Date:   Tue Oct 7 08:53:34 2014 -0400

    CI: add appveyor support

commit e1083eee34c5c5dd1d102bd2372d9d6a2ecfdfcb
Merge: 685eb0d 9bfb8ec
Author: jreback <[email protected]>
Date:   Tue Oct 7 15:57:10 2014 -0400

    Merge pull request #8504 from jreback/mpl_fix

    COMPAT: matplotlib 1.4.0 version compat (GH8502)

commit 9bfb8ec735a1d15dcd3c9804d1f12f7e33ce19e9
Author: jreback <[email protected]>
Date:   Tue Oct 7 15:15:38 2014 -0400

    COMPAT: matplotlib 1.4.0 version compat (GH8502)

commit 685eb0d73863562d56b1bf119a18709615e2739f
Merge: 2f2127f f3bcbe5
Author: jreback <[email protected]>
Date:   Tue Oct 7 13:43:04 2014 -0400

    Merge branch 'jbradish-master'

commit f3bcbe51f8b25400384cf4ff6ea3592e8fd6254e
Author: jreback <[email protected]>
Date:   Tue Oct 7 11:32:04 2014 -0400

    Merge pull request #8497 from jbradish/master

commit 2f2127f4d05889d4a53773faad57df5eaaf97c81
Merge: 29de89c d0064aa
Author: jreback <[email protected]>
Date:   Tue Oct 7 11:32:04 2014 -0400

    Merge pull request #8497 from jbradish/master

    DOC/BUG:Added Missing Semicolon to Release Note v0.15

commit d0064aae4b1e2fbe22695ec109e5d1c0d2bd93bb
Author: Joe Bradish <[email protected]>
Date:   Tue Oct 7 10:25:42 2014 -0500

    Added Missing Semicolon to Release Note v0.15

commit 29de89c1d961bea7aa030422b56b061c09255b96
Author: jreback <[email protected]>
Date:   Mon Oct 6 20:48:20 2014 -0400

    DOC: release notes update

commit d5694d9b75125c3bd1e04b691f48754dfbf0a42d
Author: jreback <[email protected]>
Date:   Mon Oct 6 20:47:44 2014 -0400

    RLS: v0.15.0rc1 IS_RELEASED=False

commit 8dfbe09c1443334fc3036465712195a36c773f4b
Author: jreback <[email protected]>
Date:   Mon Oct 6 20:38:39 2014 -0400

    RLS: 0.15.0rc1

commit 09f740b06ca70c510571f20fa6f450fd3e0cd5fe
Author: jreback <[email protected]>
Date:   Mon Oct 6 20:25:55 2014 -0400

    DOC: update release.rst for 0.15.0

commit 3d3193848bd8b88ab78e20cad5de25cd3be8d9bb
Merge: d865428 feb5d4c
Author: Joris Van den Bossche <[email protected]>
Date:   Tue Oct 7 02:15:24 2014 +0200

    Merge pull request #8330 from jorisvandenbossche/sql-2908-chunksize

    ENH: add chunksize argument to read_sql (GH2908)

commit feb5d4c7e91755016842b40f4bcc5d6f0241c614
Author: Joris Van den Bossche <[email protected]>
Date:   Sat Sep 20 17:20:23 2014 +0200

    ENH: add chunksize argument to read_sql (GH2908)

commit d865428931045837f59625cfb62d72823c5e0819
Merge: 20525f5 d573120
Author: Joris Van den Bossche <[email protected]>
Date:   Tue Oct 7 00:28:57 2014 +0200

    Merge pull request #8478 from TomAugspurger/doc-modifications

    DOC: Some more vis cleanups

commit 20525f5f57159b4056436c4d8596ed499e3deca9
Merge: 7c15827 9b94968
Author: Joris Van den Bossche <[email protected]>
Date:   Mon Oct 6 23:58:40 2014 +0200

    Merge pull request #8492 from jorisvandenbossche/fix-searchsorted

    BUG: searchsorted sorter issue with 32bit platforms (GH8490)

commit 9b94968fb3d2b91c62b984938d420fac07652a4b
Author: Joris Van den Bossche <[email protected]>
Date:   Mon Oct 6 22:20:33 2014 +0200

    BUG: searchsorted sorter issue with 32bit platforms (GH8490)

commit 7c158274d5d683f002eba03725ca83372b00df76
Merge: 5995dc3 e2bd0d3
Author: Joris Van den Bossche <[email protected]>
Date:   Mon Oct 6 22:33:17 2014 +0200

    Merge pull request #8493 from jorisvandenbossche/doc-plot-docstring

    DOC: fix notes in plot() docstring

commit e2bd0d31ccbcfcca9529dffcc258d022d6c05652
Author: Joris Van den Bossche <[email protected]>
Date:   Mon Oct 6 22:31:03 2014 +0200

    DOC: fix notes in plot() docstring

commit 5995dc36a090c193bca20d18f7089503a8674bbb
Author: Ross Petchler <[email protected]>
Date:   Mon Oct 6 10:26:40 2014 -0400

    ENH: Add orient argument and split option to DataFrame.to_dict. (GH7840)

    Update documentation with deprecation and enhancement notices.

    Remove indentation from list in docstring.

commit d65c9d7fb85f0c5e513a49ed1b94571ff44020df
Merge: 795e059 00fb38e
Author: jreback <[email protected]>
Date:   Mon Oct 6 13:09:18 2014 -0400

    Merge pull request #8488 from hunterowens/speed-docs

    DOC: warning about copying for df.append/concat. Fixes #7967

commit 00fb38ebd6acbe1f9680ebcddce94ddea6a06986
Author: hunterowens <[email protected]>
Date:   Mon Oct 6 10:59:31 2014 -0500

    DOC: warning about copying for df.append/concat. Fixes #7967

    updating concat section to reflex comments on PR #8488

    DOC: placing the concat notes in a .. note:: block. #8488

    DOC: Minor formatting changes for #8488

commit 795e05944c80e9dcdb49d7aa4550cf85a1f2a36d
Merge: 9010424 f92b2ab
Author: Joris Van den Bossche <[email protected]>
Date:   Mon Oct 6 17:09:27 2014 +0200

    Merge pull request #8487 from jorisvandenbossche/doc-fixes

    DOC: fix bunch of doc build errors

commit 9010424c3a31ad6c1bb13615e5020839d2ea3ac4
Merge: 3834259 bce77ba
Author: jreback <[email protected]>
Date:   Mon Oct 6 11:01:11 2014 -0400

    Merge pull request #8485 from jreback/index_fixups

    DOC: indexing.rst fixes w.r.t (GH8227)

commit f92b2abe8228e9899545faff0892bba61dc7e4a4
Author: Joris Van den Bossche <[email protected]>
Date:   Mon Oct 6 16:46:18 2014 +0200

    DOC: fix bunch of doc build errors

commit bce77ba83e9be04f9b7a6184d086343dff6e69d2
Author: jreback <[email protected]>
Date:   Mon Oct 6 10:27:22 2014 -0400

    DOC: indexing.rst fixes w.r.t (GH8227)

commit 3834259a93ab0b5540fe10fb7f77be647165c573
Merge: fe1a68b 5724a2a
Author: jreback <[email protected]>
Date:   Mon Oct 6 09:54:51 2014 -0400

    Merge pull request #8484 from jreback/groupby_fix

    BUG: Bug in groupby .apply with a non-affecting mutation in the function (GH8467)

commit fe1a68bc3421a49649a47f394f3c0a14f09648bb
Author: jreback <[email protected]>
Date:   Mon Oct 6 09:22:55 2014 -0400

    DOC: reference to docs for describe

commit 3d5d4465017a2e29aa975873ce73629d67ecb357
Merge: 9b024b5 45be995
Author: jreback <[email protected]>
Date:   Mon Oct 6 09:21:02 2014 -0400

    Merge pull request #8483 from jreback/value_range

    DEPR: deprecate value_range (GH8481)

commit 5724a2a83574cc6b279f13a34b2bf5882bcc5cb7
Author: jreback <[email protected]>
Date:   Mon Oct 6 09:19:07 2014 -0400

    BUG: Bug in groupby .apply with a non-affecting mutation in the function (GH8467)

commit 9b024b52caa70eb595802db9c113cffd0b1d3f11
Merge: 6d3803d 70e3ff3
Author: jreback <[email protected]>
Date:   Mon Oct 6 09:03:02 2014 -0400

    Merge pull request #8480 from behzadnouri/align-mi

    BUG: sub-frame assignment of a multi-index frame breaks alignment

commit 45be995d5ebd25398b1b6e6b2bbffdb8f44ecaf8
Author: jreback <[email protected]>
Date:   Mon Oct 6 08:46:48 2014 -0400

    DEPR: deprecate value_range (GH8481)

commit 70e3ff3bcce0b12275e3b3322c2fcf4327255ff8
Author: behzad nouri <[email protected]>
Date:   Sun Oct 5 21:28:52 2014 -0400

    sub-frame assignment of a multi-index frame breaks alignment

commit d5731209c9f038a2493dc04e465e2c31c897da74
Author: TomAugspurger <[email protected]>
Date:   Sun Oct 5 21:23:49 2014 -0500

    DOC: Fix some warnings is visualization.rst

commit 3ece879b866327c65d0f003a0ec5b8b946473b49
Author: TomAugspurger <[email protected]>
Date:   Sun Oct 5 21:23:00 2014 -0500

    DOC: Try to use matplotlib ggplot style

commit 6d3803ddcd8a7665f108a12f1f798f4078b0ad8d
Author: Benjamin Thyreau <[email protected]>
Date:   Wed Sep 3 14:35:53 2014 +0900

    More flexible describe() via include/exclude type filtering

    This enhance describe()'s output via new include/exclude list arguments,
    letting the user specify the dtypes to be summarized as output.
    This provides an simple way to overcome the automatic type-filtering done
    by default; it's also convenient with groupby().
    Also includes documentation and changelog entries.

commit 72a051c85463f5d8b32ca99616276c7e6a042e8b
Merge: d22b382 d27e37a
Author: jreback <[email protected]>
Date:   Sun Oct 5 19:52:47 2014 -0400

    Merge pull request #8476 from jreback/td_std

    BUG: allow std to work with timedeltas (GH8471)

commit d27e37af3bef8c88bc7b1d7ff7ac482cc71b9bb8
Author: jreback <[email protected]>
Date:   Sun Oct 5 16:06:45 2014 -0400

    BUG: allow std to work with timedeltas (GH8471)

commit d22b382753d07ecba7346afa2779e0af78366614
Merge: 54621c7 93fd023
Author: Joris Van den Bossche <[email protected]>
Date:   Sun Oct 5 23:24:57 2014 +0200

    Merge pull request #8474 from jorisvandenbossche/sql-8314-time

    TEST: add for sqlite fallback raising on datetime.time (and not failing silently) GH8341

commit 54621c73b7df40503c0d00ece250159a75b4b1f9
Merge: c741a54 8c51d86
Author: jreback <[email protected]>
Date:   Sun Oct 5 17:15:45 2014 -0400

    Merge pull request #8475 from jreback/at_fix

    BUG: .at indexing should allow enlargement scalars w/o regards to the type of index (GH8473)

commit 8c51d868ed19dcd8ed346cd4f61e7494eb18aa05
Author: jreback <[email protected]>
Date:   Sun Oct 5 16:41:09 2014 -0400

    BUG: .at indexing should allow enlargement scalars w/o regards to the type of index (GH8473)

commit 93fd0231e36a15cdb91541c8de4024c8dd2aef33
Author: Joris Van den Bossche <[email protected]>
Date:   Sun Oct 5 22:03:23 2014 +0200

    TEST: add for sqlite fallback raising on datetime.time (and not failing silently) GH8341

commit c741a5471f683db7ec793510cd79239308750e11
Author: Joris Van den Bossche <[email protected]>
Date:   Sun Oct 5 22:11:53 2014 +0200

    DOC: fix minor formatting issues

commit 4e439d97709ab3b0b78d7c93d4962317a31f605e
Merge: 0e11eea 55ba0f0
Author: Joris Van den Bossche <[email protected]>
Date:   Sun Oct 5 21:29:00 2014 +0200

    Merge pull request #8440 from jorisvandenbossche/sql-ooapi-naming

    API: SQL class definitions renaming

commit 0e11eea1499a97861f97169b0efe18daa4f8f705
Merge: c350247 913ece8
Author: Joris Van den Bossche <[email protected]>
Date:   Sun Oct 5 20:01:12 2014 +0200

    Merge pull request #8470 from jorisvandenbossche/sql-8314-time

    SQL/ERR: raise error when insert failed with sqlite fallback (GH8341)

commit c3502477450deac8dfb973efa2c1914d3d4fa865
Merge: bc6a1f7 722e4ef
Author: jreback <[email protected]>
Date:   Sun Oct 5 13:03:01 2014 -0400

    Merge pull request #8472 from rockg/series-indexing

    Disallow Series indexing by DataFrame

commit bc6a1f79963bdee03b0fd770a80b013c56743da3
Author: jreback <[email protected]>
Date:   Sun Oct 5 11:40:45 2014 -0400

    TST: dtype fixes on tests wrt GH8432

commit 722e4efd99d28590b9c283fd0d91b6e405846088
Author: rockg <[email protected]>
Date:   Sun Oct 5 11:14:09 2014 -0400

    Disallow Series indexing by DataFrame

commit 65d716ce0ae992c09fca2142b78e8432fc4f6646
Author: jreback <[email protected]>
Date:   Sun Oct 5 10:19:00 2014 -0400

    DOC: v0.15.0 / faq.rst updates

commit d29d4c6623a29db36a4eb8d062bdcfe603e2b490
Author: jreback <[email protected]>
Date:   Sun Oct 5 09:45:41 2014 -0400

    COMPAT: disallow Timedelta invalid ops

commit bca569903ec52a28c0e13e9f4a9f494abd22077e
Merge: d1c0271 8980295
Author: jreback <[email protected]>
Date:   Sun Oct 5 10:05:18 2014 -0400

    Merge pull request #8462 from immerrr/preserve-index-type-when-reindexing-to-empty

    BUG: fix Index.reindex to preserve type when target is empty list/ndarray

commit 913ece8c65757c045a9f15aa8210a40c48a80ff6
Author: Joris Van den Bossche <[email protected]>
Date:   Sun Oct 5 12:04:47 2014 +0200

    SQL/ERR: raise error when insert failed with sqlite fallback (GH8341)

commit 8980295a42ff8f1ae1889e36c69970ab25338f13
Author: immerrr <[email protected]>
Date:   Sat Oct 4 19:44:28 2014 +0400

    BUG: fix Index.reindex to preserve type when target is empty list/ndarray

    TST: check index/columns types when doing empty loc/ix tests

    CLN: don't _ensure_index in NDFrame._reindex_axes, it is done in
         Index.reindex

commit d1c02710f5e7de6388489269c9cd450e9b94b6d3
Merge: 5a3af87 ab97e0a
Author: jreback <[email protected]>
Date:   Sat Oct 4 22:31:10 2014 -0400

    Merge pull request #7619 from asobrien/df-mem-info

    ENH: dataframe memory usage

commit ab97e0a3fa542416fc28db3c55cea4f3ce640f19
Author: Anthony O'Brien <[email protected]>
Date:   Sat Oct 4 20:54:34 2014 -0500

    Closes #6852. Dataframe memory usage methods implemented.

commit 5a3af87b980ef5d009b21c5b61c201a640d95b12
Merge: 8128b0b b766a74
Author: Phillip Cloud <[email protected]>
Date:   Sat Oct 4 21:02:29 2014 -0400

    Merge pull request #7447 from cpcloud/pd-searchsorted

    WIP: searchsorted implementation

commit 8128b0bf8756c17ad1c622d9d2ce691e7e6433a7
Author: jreback <[email protected]>
Date:   Sat Oct 4 20:29:55 2014 -0400

    TST: complex ops test correction

commit b766a74fdd07749eeb71672a9cf09674bc881ae1
Author: Phillip Cloud <[email protected]>
Date:   Thu Jun 12 23:02:49 2014 -0400

    WIP: searchsorted implementation

commit 02161ffa2f028830e408195e0e9157e217a18235
Merge: ba4f146 c1e3369
Author: jreback <[email protected]>
Date:   Sat Oct 4 16:02:08 2014 -0400

    Merge pull request #8460 from immerrr/preserve-index-names-when-reindexing

    BUG: fix Index.reindex to preserve name when target is list/ndarray

commit c1e3369f83b0b759a5c91d58637829d16023a01d
Author: immerrr <[email protected]>
Date:   Sat Oct 4 11:10:05 2014 +0400

    BUG: fix Index.reindex to preserve name when target is list/ndarray

    CLN: drop copy_if_needed kwarg of Index.reindex, it's True everywhere

    TST: enable back empty loc/ix tests that failed before

    DOC: Bump Index.reindex docstring

commit ba4f146ee120e8afa0b093060f4ade662352dc10
Author: jreback <[email protected]>
Date:   Sat Oct 4 15:02:23 2014 -0400

    DOC: categorical.rst fixups

commit 0612230ef26f46535d4729e7204438bec2b72b23
Author: jreback <[email protected]>
Date:   Sat Oct 4 13:23:09 2014 -0400

    TST/PERF: have moments slow tests run faster (GH8450)

commit daf87612c278dd50177b32436374c393cbb3305d
Author: jreback <[email protected]>
Date:   Sat Oct 4 14:48:52 2014 -0400

    DOC: doc updates for categoricals

commit 1ac2dec8f4dd2a75c5fde394cc9c9ea2de5c51a2
Merge: a0bef0b 3810417
Author: Tom Augspurger <[email protected]>
Date:   Sat Oct 4 13:01:29 2014 -0500

    Merge pull request #8391 from TomAugspurger/square-pie

    DOC: Use square figsize for pie plots

commit a0bef0bfd5ebd28997ac35133822bb097b6e90db
Merge: 9872f31 30246a7
Author: jreback <[email protected]>
Date:   Sat Oct 4 13:19:50 2014 -0400

    Merge pull request #8457 from behzadnouri/align-rhs

    BUG: type diversity breaks alignment

commit 9872f31b18a493ed53bb636869254b281986c694
Merge: a1fc667 0b12eaa
Author: jreback <[email protected]>
Date:   Sat Oct 4 12:38:15 2014 -0400

    Merge pull request #8455 from jreback/cat2

    PERF: optimize storage type for codes in Categoricals (GH8453)

commit a1fc66786fb05f2b8bab394d17e3e441f7a78f4e
Merge: 8560bfa e90474a
Author: jreback <[email protected]>
Date:   Sat Oct 4 12:20:49 2014 -0400

    Merge pull request #8461 from jreback/applymap

    BUG: fix applymap to handle Timedelta

commit 3810417de527708fd7885580798c4dcdb9ab6134
Author: TomAugspurger <[email protected]>
Date:   Thu Sep 25 12:06:22 2014 -0500

    DOC: Cleanup of plots in visualization.py

    DOC: Use square figsize for pie plots
    DOC: Don't share axis labels and reorder to fit in page.

commit 8560bfa33de84d3cccb0617839aa1ac19d4ac2f9
Merge: d595a6c 678409f
Author: Tom Augspurger <[email protected]>
Date:   Sat Oct 4 10:09:16 2014 -0500

    Merge pull request #8459 from TomAugspurger/plot_cln

    VIS: Cleanups in plotting.py

commit e90474a506a05281974f93f4153b513080e807b3
Author: jreback <[email protected]>
Date:   Sat Oct 4 10:49:25 2014 -0400

    BUG: fix applymap to handle Timedelta

commit 30246a7d2b113b7115fd0708655ae8e1c3f2e6c5
Author: behzad nouri <[email protected]>
Date:   Fri Oct 3 18:02:39 2014 -0400

    type diversity breaks alignment

commit 0b12eaa6fa025b9f70c06fe0a1bb436585d51218
Author: jreback <[email protected]>
Date:   Fri Oct 3 16:51:27 2014 -0400

    PERF: optimize storage type for codes in Categoricals (GH8453)

commit d595a6cb53c50c7735f37d3573714165640872da
Merge: f7426d6 6348786
Author: Tom Augspurger <[email protected]>
Date:   Sat Oct 4 09:09:10 2014 -0500

    Merge pull request #8429 from TomAugspurger/flier-test-adjust

    TST: Adjust boxplot tests following MPL API change

commit f7426d615d8801eaaf4f104cf71bfbd7eb9d65ce
Author: jreback <[email protected]>
Date:   Sat Oct 4 09:41:30 2014 -0400

    TST: fix path for getting pickles wrt (GH8431)

commit b84d237e742412d91463bd07ed3906f9085d7440
Merge: 46df5bb 68f6268
Author: jreback <[email protected]>
Date:   Sat Oct 4 09:37:55 2014 -0400

    Merge pull request #8293 from bjonen/fix_8236_2

    PERF: Improve performance CustmBusinessDay - 2nd

commit 678409fb1e9698dfa55f5380b936d1ffe5b12fc5
Author: TomAugspurger <[email protected]>
Date:   Sat Oct 4 08:28:37 2014 -0500

    PEP8 fixups

commit 46df5bb8a086c615fe78669c31f49ecac53b811c
Author: jreback <[email protected]>
Date:   Sat Oct 4 09:17:13 2014 -0400

    TST: compat for *.pkl -> *.pickle in pandas/tests/data

commit 55ba0f065997dea788b05feab854818f1850845b
Author: Joris Van den Bossche <[email protected]>
Date:   Thu Oct 2 10:25:23 2014 +0200

    API: SQL class definitions renaming

    - renaming of PandasSQLAlchemy to SQLDatabase and PandasSQLTable to SQLTable
    - renaming of SQLDatabase.read_sql to read_query (to be consistent with read_query/read_table)
    - legacy -> fallback

commit 68f626872676604f34190580df23d9653a389e03
Author: bjonen <[email protected]>
Date:   Tue Sep 16 00:37:37 2014 +0200

    PERF: Improve performance of CustomBusinessDay

commit 53cc1b7447ced5fde96d887af414f801b2913494
Author: sinhrks <[email protected]>
Date:   Sun Jul 27 14:29:26 2014 +0900

    DOC/CLN: Cleanups plotting.py

commit fccd7feabfe0bf15c30d712f3a4a4ff71e76ad4a
Merge: 29b1aa7 6479766
Author: jreback <[email protected]>
Date:   Fri Oct 3 15:12:07 2014 -0400

    Merge pull request #8454 from jreback/legacy_pickle

    BUG: reset identity on legacy index pickles (GH8431)

commit 64797669fbfff2698911d541bb2bfbc774f686af
Author: jreback <[email protected]>
Date:   Fri Oct 3 14:41:20 2014 -0400

    BUG: reset identity on legacy index pickles (GH8431)

commit 29b1aa7a9904a4819e2da95171e94cd55fbab194
Merge: 0e69f9e 115596b
Author: jreback <[email protected]>
Date:   Fri Oct 3 14:00:03 2014 -0400

    Merge pull request #8451 from jsexauer/fix_delevel

    Remove DataFrame.delevel

commit 115596bf5fbc5f0f11ff0c02531872b24373ec9b
Author: unknown <[email protected]>
Date:   Fri Oct 3 11:07:05 2014 -0400

    Remove DataFrame.delevel

commit 0e69f9e4d6d1a2b718c90c9828227d9945184cd4
Merge: 977034a 8a90c7b
Author: jreback <[email protected]>
Date:   Fri Oct 3 08:33:46 2014 -0400

    Merge pull request #8449 from jreback/cat

    BUG: bug in df.info() when embedded categorical (related GH7619)

commit 8a90c7b3dc524ef0fed375556fa1040dd5fe28f8
Author: jreback <[email protected]>
Date:   Fri Oct 3 08:04:36 2014 -0400

    BUG: bug in df.info() when embedded categorical (related GH7619)

commit 977034a1268575ed75bcf3103cee43048673c347
Merge: f8582e8 1ff4b36
Author: Phillip Cloud <[email protected]>
Date:   Thu Oct 2 13:02:28 2014 -0400

    Merge pull request #7971 from cpcloud/conda

    CI/WIP: Use conda for most deps

commit 1ff4b36eefcb693c84017ac588c67e792fafd1d7
Author: Phillip Cloud <[email protected]>
Date:   Wed Jul 9 16:04:40 2014 -0400

    CI: use conda for most build dependencies

commit f8582e8cffba95e06301d41a08437cdd0bae2707
Merge: 2b389b4 0d8c7fd
Author: jreback <[email protected]>
Date:   Thu Oct 2 11:15:11 2014 -0400

    Merge pull request #8443 from unutbu/dtypeO-equals

    BUG: NDFrame.equals gives false negatives with dtype=object (GH8437)

commit 0d8c7fd91581616749b35f7a54e3f302147b1fb5
Author: unutbu <[email protected]>
Date:   Thu Oct 2 09:39:49 2014 -0400

    BUG: NDFrame.equals gives false negatives with dtype=object (GH8437)

commit 2b389b41f82113ff30bfc9fa55bac1c24088f22d
Merge: 01d275f 4b443f9
Author: Joris Van den Bossche <[email protected]>
Date:   Thu Oct 2 08:41:41 2014 +0200

    Merge pull request #8439 from bthyreau/doc_selectdtypes_categ

    DOC: mention 'category' in select_dtypes docstring

commit 4b443f9db40597e6d5ddbcb3dfb8079803a4a325
Author: Benjamin Thyreau <[email protected]>
Date:   Thu Oct 2 10:46:00 2014 +0900

    DOC: mention 'category' in select_dtypes docstring

commit 01d275fb9cb9bc724675bd6cd9420df65f1d80da
Merge: 8d3a39a b3ce297
Author: jreback <[email protected]>
Date:   Wed Oct 1 09:03:57 2014 -0400

    Merge pull request #8434 from jreback/bin_grouper

    BUG: Groupby.transform related to BinGrouper and GH8046 (GH8430)

commit b3ce297fe8f388b05fca2739535eb9717c8a04d4
Author: jreback <[email protected]>
Date:   Wed Oct 1 08:33:28 2014 -0400

    BUG: Groupby.transform related to BinGrouper and GH8046 (GH8430)

commit 6348786067e112ca371a02c722401309459a039f
Author: TomAugspurger <[email protected]>
Date:   Tue Sep 30 14:35:34 2014 -0500

    TST: Adjust boxplot tests following MPL API change

    fliers on a boxplot are now 1 object instead of 2.

commit 8d3a39a220d209f5901ae652b37749f5abdf4c25
Merge: 4074957 4802d0f
Author: jreback <[email protected]>
Date:   Tue Sep 30 15:39:45 2014 -0400

    Merge pull request #8417 from jreback/mi_grouping

    BUG: regression in groupby with a pass thru multiindex on axis=1 (GH7997)

commit 4802d0f2177d0c96618ce4a54f8a9e68fe23bad8
Author: jreback <[email protected]>
Date:   Mon Sep 29 11:48:54 2014 -0400

    BUG: regression in groupby with a pass thru multiindex on axis=1 (GH7997)

commit 4074957c0933893afcb2ec785fcba6520789c1df
Author: RenzoBertocchi <[email protected]>
Date:   Mon Sep 22 17:17:50 2014 +0200

    BUG: Bug on pivot_table with margins and dict aggfunc (GH8349)

commit c2135239aede02d958499770b3c6eb8f330362dc
Merge: e871245 c494c03
Author: jreback <[email protected]>
Date:   Tue Sep 30 11:48:30 2014 -0400

    Merge branch 'vbench_for_groupby' of https://github.com/dlovell/pandas into dlovell-vbench_for_groupby

commit e8712455bb4a2f4e509a38c2d3ba386f3b9d4e72
Merge: 5cfc9cf a95d84a
Author: jreback <[email protected]>
Date:   Tue Sep 30 10:10:34 2014 -0400

    Merge pull request #8423 from jreback/str

    DOC: create text.rst with string methods (GH8416)

commit a95d84a79875e795ca7bc59d5f904eaed0c26e71
Author: Dav Clark <[email protected]>
Date:   Mon Sep 29 19:11:34 2014 -0700

    Some more clarification, particularly regexps

commit c494c03165ae1e1195893d6567729327021bb8ae
Author: dlovell <[email protected]>
Date:   Tue Sep 30 07:21:49 2014 -0400

    make running multiple values of ngroups easy

commit fe924b440c2403399c4f274429ff88a4f113a470
Author: jreback <[email protected]>
Date:   Mon Sep 29 17:50:13 2014 -0400

    DOC: create text.rst with string methods (GH8416)

commit 5cfc9cfd408e7496a935b821d3a6e8a41a481d7a
Merge: cca1aab 3f26ee6
Author: jreback <[email protected]>
Date:   Mon Sep 29 17:43:23 2014 -0400

    Merge pull request #8415 from jreback/tz_aware

    API/BUG: a UTC object inserted into a Series/DataFrame will preserve the UTC and be object dtype (GH8411)

commit cca1aab4eecd23764ab3d6a46cfe9a58f4e84eac
Merge: 0f83679 70f11e4
Author: jreback <[email protected]>
Date:   Mon Sep 29 17:42:44 2014 -0400

    Merge pull request #8400 from benjschiller/master

    add support for numpy 1.8+ data types for conversion to r dataframe

commit 0f8367957e8ee5326416f79e2f70d3aeec2fb69c
Merge: 170cf12 5fc1daa
Author: jreback <[email protected]>
Date:   Mon Sep 29 17:36:42 2014 -0400

    Merge pull request #8413 from JanSchulz/CategoricalFixups3

    Categorical doc fixups

commit 170cf12c84d2a6587b35e7d462dd65067e0bc5eb
Merge: 863c891 c011ee5
Author: jreback <[email protected]>
Date:   Mon Sep 29 16:55:13 2014 -0400

    Merge pull request #8418 from cstoafer/docs

    Moved startup script information to options docs and fixed link

commit 863c8914a6f2697f905fea4c5e2459b640d15496
Merge: 1aa7c2c d93f464
Author: jreback <[email protected]>
Date:   Mon Sep 29 16:52:02 2014 -0400

    Merge pull request #8421 from JanSchulz/cat_searchsorted

    FIX: Add Categorical.searchsorted stub

commit 70f11e4137c3f749605d3612d537e48536eb1a60
Author: Ben Schiller <[email protected]>
Date:   Fri Sep 26 16:32:48 2014 -0700

    Fix VECTOR_TYPES and NA_TYPES dicts used for pd DataFrame <-> R data.frame conversion

    added commas to dict

    string_ not string

    changed parse_version to LooseVersion

    Updated release notes

commit d93f46447531c5983b05687b268f3983776459d4
Author: Jan Schulz <[email protected]>
Date:   Mon Sep 29 20:03:36 2014 +0200

    FIX: Add Categorical.searchsorted stub

    For https://github.com/pydata/pandas/pull/7447, add a searchsorted
    stub, which simple raises NotImplementedError, so taht we raise a
    more clear error than attribute not found.

commit 5fc1daa1d7163e63cb20bd0ae23e7e4589e2dbe4
Author: Jan Schulz <[email protected]>
Date:   Mon Sep 29 16:01:03 2014 +0200

    Doc: Add Categorical to R comparison

    Also fix all omited `pd.`

commit c011ee5b0b5b851a72d43ddb24ff18c9a6864d8e
Author: Chris Stoafer <[email protected]>
Date:   Mon Sep 29 12:19:17 2014 -0400

    Moved startup script information to options docs and fixed link based on GH8407 and GH5748

commit 1aa7c2c94170f5a0fd0daf3c90cb89f1aa85d8c3
Author: behzad nouri <[email protected]>
Date:   Thu Sep 25 23:17:51 2014 -0400

    BUG: apply Series mask to aligned new values (GH8387)

commit 9b46a6980cf89a8a64379c881887a9165bd23498
Merge: 68fdacd 26905ab
Author: jreback <[email protected]>
Date:   Mon Sep 29 10:52:54 2014 -0400

    Merge pull request #8405 from sh9189/bbopensrcday

    Better message in exception when conversion from period to timestamp fai...

commit 3f26ee6f0818abfe6bf01608f9d88e19e2865f7c
Author: jreback <[email protected]>
Date:   Mon Sep 29 10:02:14 2014 -0400

    API/BUG: a  timeseries/index localized to UTC when inserted into a Series/DataFrame will preserve the UTC timezone (rather than being a naive datetime64[ns]) as object dtype (GH8411)

commit 68fdacd12e7963706f0ac1b859058bddb6b814bd
Author: tom-alcorn <[email protected]>
Date:   Tue Sep 23 14:59:00 2014 -0400

    BUG: Bug in interesection of non-monotonic non-unique indexes (GH8362)

commit b8d680dd8a91f3066027112e9e1c65ac1feeb1dd
Merge: dbcf18a 6e8ce68
Author: jreback <[email protected]>
Date:   Mon Sep 29 10:05:45 2014 -0400

    Merge pull request #8414 from JanSchulz/nbytes_categories

    FIX: add nbytes property in Categorical

commit 65b1d918305a836360ea2ae16d6ef1906031bfee
Author: Jan Schulz <[email protected]>
Date:   Mon Sep 29 15:44:15 2014 +0200

    Fixup 10min

commit dbcf18aa50228d4645a8f3f0fbed45375b9a817c
Merge: 0c2c2ef 27afaeb
Author: jreback <[email protected]>
Date:   Mon Sep 29 09:23:48 2014 -0400

    Merge pull request #8409 from jacobschaer/issue-8107

    Allow importing pandas without setuptools

commit 0c2c2ef4b52969fe5ad16014db3181bf38af978d
Merge: e487a30 b9bc371
Author: jreback <[email protected]>
Date:   Mon Sep 29 09:21:49 2014 -0400

    Merge pull request #8406 from sh9189/bbopensrcday2

    CLN: Remove core/array.py

commit 6e8ce685eb5a4bd0b39665a3a0d7ddd627ea8ed0
Author: Jan Schulz <[email protected]>
Date:   Mon Sep 29 15:10:08 2014 +0200

    FIX: add nbytes property in Categorical

commit 009097f41a1088062bb04909adee648a7e946b34
Author: Jan Schulz <[email protected]>
Date:   Mon Sep 29 15:00:35 2014 +0200

    DOC: update cut/qcut docstrings

    * add some clarification about the useage of bins/labels and that they
      end up as categories.
    * Update the examples to produce the current output with the new
      categoricals.

commit 66cdd6674b897939ec2e8bc095f55ef682123849
Author: Jan Schulz <[email protected]>
Date:   Sun Sep 28 21:58:39 2014 +0200

    DOC: fixup categorical news

commit 3d5db57c76781ae3e2aeaa43e511566d71694107
Author: dlovell <[email protected]>
Date:   Sun Sep 28 13:47:35 2014 -0400

    BENCH: programmatically create benchmarks for large ngroups (GH6787)

commit 27afaeb51ab87d49dd451f12e7f3d009ef00f6cb
Author: Jacob Schaer <[email protected]>
Date:   Sat Sep 27 14:09:00 2014 -0700

    Allow importing pandas without setuptools

commit e487a3018549a8a222c17cb4996e55a80d756463
Merge: 9b63bfc f80b7fe
Author: jreback <[email protected]>
Date:   Sat Sep 27 17:06:30 2014 -0400

    Merge pull request #8407 from cstoafer/docs

    Added a section to FAQ Docs about startup scripts for setting up pandas based on issue #5748.

commit f80b7fe57ab0f2e046ff0ea7c67ec3635f274024
Author: Chris Stoafer <[email protected]>
Date:   Sat Sep 27 16:03:23 2014 -0400

    Added a section to FAQ Docs about startup scripts for setting up pandas.

commit b9bc3719e9224e0d4cf72556d2284339612cca61
Author: Shahul Hameed <[email protected]>
Date:   Sat Sep 27 15:34:50 2014 -0400

    CLN: Remove core/array.py

commit 26905ab9cba6591fef7e0073214e80527dc6b52b
Author: Shahul Hameed <[email protected]>
Date:   Sat Sep 27 13:57:59 2014 -0400

    Better message in exception when conversion from period to timestamp fails

commit 9b63bfc2c9bc18ab96ddc9dda293973d180300d4
Merge: d3976f5 6dc9eb9
Author: Tom Augspurger <[email protected]>
Date:   Sat Sep 27 08:08:26 2014 -0500

    Merge pull request #8398 from TomAugspurger/mpl-dev-boxplot-skips

    TST: Skip failing boxplot tests on mpl 1.4+

commit d3976f59189a11bc912ced468715bfd15133f9fb
Merge: b296b9e 52aee10
Author: jreback <[email protected]>
Date:   Fri Sep 26 21:48:08 2014 -0400

    Merge pull request #8357 from amras1/new-float-conversion

    Changed round_trip converter to Python converter instead of strtod

commit b296b9e9164efc2f6bdbd96eb361415336acdc3f
Author: jreback <[email protected]>
Date:   Fri Sep 26 19:57:20 2014 -0400

    DOC: v0.15.0 et all updates

commit adcdb5b92882712066883e608daeae850930bb91
Merge: cbf3679 6c3f184
Author: jreback <[email protected]>
Date:   Fri Sep 26 19:36:40 2014 -0400

    Merge pull request #8399 from jreback/panel

    BUG: inconsisten panel indexing with aligning frame (GH7763)

commit 52aee10acc390fc230eead4287e3341692a3500a
Author: Michael Mueller <[email protected]>
Date:   Mon Sep 22 09:31:47 2014 -0400

    Changed round_trip converter to Python converter instead of strtod

commit cbf367917ddb6e5ce1c4e9a4c6d7196166c111e4
Merge: 62ac12b 04f3d40
Author: jreback <[email protected]>
Date:   Fri Sep 26 18:06:18 2014 -0400

    Merge pull request #8373 from jbradish/master

    ENH: Added file path existence check for read_hdf

commit 6c3f1848a8948127957e4d12d1bee40a8866283f
Author: jreback <[email protected]>
Date:   Thu Sep 18 08:59:48 2014 -0400

    BUG: inconsisten panel indexing with aligning frame (GH7763)

commit 04f3d400fc196e4bf37583d9216ab9f83f5ce39d
Author: Joe Bradish <[email protected]>
Date:   Fri Sep 26 13:37:00 2014 -0500

    Added file path existence check for read_hdf

commit 62ac12b15a6375dbb58e785b68c3cc9a601ee008
Merge: 68e5c59 a638a4f
Author: jreback <[email protected]>
Date:   Fri Sep 26 11:57:37 2014 -0400

    Merge pull request #8372 from jreback/resample

    BUG: bug in non-evently divisible offsets when resampling (e.g. '7s') (GH8371)

commit 68e5c591c7459350001b69e9619bd20e27d39e2f
Merge: df3224f 8818a2c
Author: jreback <[email protected]>
Date:   Fri Sep 26 11:19:39 2014 -0400

    Merge pull request #8388 from unutbu/fillna-df

    EHN: Allow DataFrame.fillna to accept a DataFrame as its fill value.

commit 6dc9eb952b371cea74ed31990a7b4162dc102ff7
Author: TomAugspurger <[email protected]>
Date:   Fri Sep 26 09:59:59 2014 -0500

    TST: Skip failing boxplot tests on mpl 1.4+

commit df3224ffa05791179d6e3c546e1f917a9a2a81f0
Author: jreback <[email protected]>
Date:   Fri Sep 26 10:32:39 2014 -0400

    FIX: categorical fixups

commit 6cf0a4e59f3aae7dafbc85b2748effa09bcef049
Merge: 0ffef1d d7f6021
Author: jreback <[email protected]>
Date:   Fri Sep 26 10:35:18 2014 -0400

    Merge pull request #8384 from mcjcode/master

    non-existent columns in DataFrame.dropna subset argument now raises KeyError

commit 0ffef1dd5aad5efd7a2b734fc21c3d8284d946fc
Merge: ec331f0 3f282da
Author: jreback <[email protected]>
Date:   Thu Sep 25 18:14:33 2014 -0400

    Merge pull request #8153 from JanSchulz/categorical_fixups_2

    Fixups for categoricals

commit ec331f0679103e9cec7a9209932b128a4b6a6d34
Merge: 1f504fe 1ab045d
Author: Tom Augspurger <[email protected]>
Date:   Thu Sep 25 14:46:38 2014 -0500

    Merge pull request #8393 from TomAugspurger/pie-nan-fu

    TST: Fix failing pie plot tests

commit 1ab045dd55cc068e80d0276c02dcba047a0287a7
Author: TomAugspurger <[email protected]>
Date:   Thu Sep 25 13:40:21 2014 -0500

    TST: Fix failing pie plot tests

commit 3f282da6aac82fabf0749933fe8259a8bd6f33a6
Author: Jan Schulz <[email protected]>
Date:   Mon Sep 22 23:50:59 2014 +0200

    DOC: update to reflect changes in Categorical

    * rename of levels -> categories
    * don't use the `Categorical` constructor if thats not needed
    * introduce the new methods to manipulate categories

commit 4c45d7f05eee2933b8271fd49d8bafb14f9fa546
Author: Jan Schulz <[email protected]>
Date:   Tue Sep 2 13:56:21 2014 +0200

    ENH: Introduce methods for each change to Categorical.categories

    There are multiple "actions" which can be done to categories:

    - add one or more categories
    - remove one or more categories
    - rename the categories
    - reorder the categories
    - remove unused categories

    Each of these actions now gets it's own method together with a special
    `set_categories`, which can do more than one action in one call, which is
    needed to save computation time.

    See https://github.com/pydata/pandas/pull/8153#issuecomment-54356611 and the
    following comments.

commit 9d8624726a23aacb9bce788051f47053a317d490
Author: Jan Schulz <[email protected]>
Date:   Sat Sep 13 18:57:22 2014 +0200

    API: rename Categorical.levels to .categories

    The name 'levels' was already used by a much different concept in
    MultiIndex and this was too confusing, so change the name to
    'categories'. Add deprecation warning if the old name is used
    (in constructor or the public 'levels' attribute). The old name
    'levels' is not anymore exposed unter Series.cat as it was never
    part of a stable release.

    See the discussion in https://github.com/pydata/pandas/issues/8074

    This rename was done by search&replace in categorical.py and the
    corresponding tests_categorical.py, implementing the deprecation
    accessor (with a temporary 'raise Exception') and then run the the
    unittests and change code until all tests pass.

commit 1f504feecf2bc3dc9c571b107310e979e19180ce
Merge: d1faa4a bc9e808
Author: Tom Augspurger <[email protected]>
Date:   Thu Sep 25 11:55:01 2014 -0500

    Merge pull request #8307 from TomAugspurger/pie-nan

    VIS: Hide labels for NaN/zeros in boxplt [WIP]

commit b74e0a3355a1adfc4140357db7e8e093add34968
Author: Jan Schulz <[email protected]>
Date:   Mon Sep 1 15:45:31 2014 +0200

    API: Implement Series.cat.codes and rename the proxy class

    * Series should also expose Series.cat.codes as a Series of codes.
      See discussion in
      https://github.com/pydata/pandas/issues/8074#issuecomment-53992337

    * Rename CategoricalProperties CategoricalAccessor. It has more
      than properties, so "accessor" is more appropiate.

commit 8fd654cab54e2071c3703639da47b70aef7342cf
Author: Jan Schulz <[email protected]>
Date:   Mon Sep 1 16:32:26 2014 +0200

    API: remove compat keyword in Categorical constructor

    Before the new Categorical work, the default two argument constructor
    was expecting "codes and levels" but this was changed to
    "values and levels" and a 'compat' kwarg was added, which could
    be used to switch to the old style constructor useage.

    It was decided that we switch to the new style constructor useage as
    the new default (compat=False) and implement a 'from_codes(...)'
    constructor.

    As the compat codepath is now never triggered and code should be
    changed to use the new 'from_codes()' constructor, remove the old
    codepath.

    Add some warnings if we are pretty sure that the old style
    constructor is meant, but unfortunately we don't catch all cases.

commit 206fb972bcd1957334126acd29862047fe9ae4d7
Author: Jan Schulz <[email protected]>
Date:   Mon Sep 1 15:20:42 2014 +0200

    FIX: Implement small Categorical fixups

    This is a squashed commit which adresses several small issues
    identified in the latest code review.

    * As per dicussion in https://github.com/pydata/pandas/issues/8074,
      Categorical.unique() should return an array and not a Index object
      to make is consistant with other unique() implementations.

    * Mark a few more categorical methods as internal

    * Change reorder_levels and drop_unused_levels to inplace=False. As per
      the discussion in #8074, change the default of both methods
      to return a copy and only do the change inplace if told otherwise.

      Discusion in
      https://github.com/pydata/pandas/issues/8074#issuecomment-54093664

    * Make str(cat) more array like. As Categorical is more like
      np.ndarray and less similar to pd.Series, change the string
      representation of Categorical to be more like np.ndarray.

      The new str(cat) will only show at max 10 values.

      Also remove the "Categorical(" in front of an empty categorical
      -> either that's put in front of every str(cta) or none. It also missed
      the final ")".

    * Test more nan handling in Categorical

    * Add a testcase for groupby with two columns and unused categories

      During the devleopment of pandas new Categorical support, groupby
      with two columns (one worked) didn't include empty categories which
      omits rows with NaN. Add a testcase so this is checked.

      No codechanges nessesary, this bug was fixed at some other places
      before. See #8138 for more details.

commit 8818a2c23ea941f3f77ace9bad884240077d0444
Author: unutbu <[email protected]>
Date:   Thu Sep 25 09:41:55 2014 -0400

    EHN: Allow DataFrame.fillna to accept a DataFrame as its fill value.

commit d7f602153a9ee43abd7b1a5719be135640a7d054
Author: mcjcode <[email protected]>
Date:   Wed Sep 24 20:24:56 2014 -0400

    non-existent columns in DataFrame.dropna subset argument now raises KeyError

commit d1faa4a808de3571e46c3001f53c64deab62ab29
Merge: 7f034bc 2ff8646
Author: jreback <[email protected]>
Date:   Wed Sep 24 07:59:41 2014 -0400

    Merge pull request #8378 from jreback/fillna_err

    API: .fillna will now raise a NotImplementedError when passed a DataFrame (GH8377)

commit 2ff86467a2dcbec86063ca55d93eca303b9926dd
Author: jreback <[email protected]>
Date:   Tue Sep 23 19:23:24 2014 -0400

    API: .fillna will now raise a NotImplementedError when passed a DataFrame (GH8377)

commit a638a4f328bc16903c4c7d7cc37af9dbdb8ada74
Author: jreback <[email protected]>
Date:   Tue Sep 23 13:59:33 2014 -0400

    BUG: bug in non-evently divisible offsets when resampling (e.g. '7s') (GH8371)

commit 7f034bcb934981b2e7c22a27f2606b3f0d225069
Author: jreback <[email protected]>
Date:   Tue Sep 23 18:51:34 2014 -0400

    TST: 32-bit fix related GH8364

commit ca2e608c22a15d49dc290ec5c645c64b95c0031d
Author: Jan Schulz <[email protected]>
Date:   Mon Sep 1 14:31:41 2014 +0200

    API/FIX: wrap result cut/qut in Series and make it ordered

    As per the discussion in #8077, the result of cut/qcut should be
    wrapped into a Series, now that this is possible.

    Also change the returned Categorical to "ordered=True", this was
    overlooked when this was introduced in Categorical.

    Closes: #8077

commit aba7d9ce460513a1c89f336c1637a546acd2b2ad
Merge: 41b1f65 39a31d1
Author: jreback <[email protected]>
Date:   Tue Sep 23 11:35:39 2014 -0400

    Merge pull request #8299 from seth-p/test_moments_warnings

    TST: suppress gratuitous warnings in test_moments.py

commit 41b1f65c0c149bab934edc5d4d12b9215713f5cc
Merge: 8a16900 cb8a7f0
Author: jreback <[email protected]>
Date:   Tue Sep 23 10:58:45 2014 -0400

    Merge pull request #8370 from jreback/tz_pickle

    BUG/COMPAT: set tz on DatetimeIndex on pickle reconstruction (GH8367)

commit 8a169001b95f25095b619f5e0770fe15005909ee
Author: klonuo <[email protected]>
Date:   Sat Sep 20 00:26:43 2014 +0200

    BUG: to_clipboard output formatting (GH8305)

commit 391e90b9939d8e912c3be16343b493c20f6ca55b
Author: Jesse Farnham <[email protected]>
Date:   Mon Sep 22 14:51:56 2014 -0400

    BUG: OLS with clustering and nw_lags does not error (GH5884)

    Added statement to raise ValueError when OLS is run with clustering and
    Newey-West adjustments. Closes GH5884.

commit cb8a7f0c69aaed9003cda5d4a5d1777b4444f799
Author: jreback <[email protected]>
Date:   Tue Sep 23 10:16:08 2014 -0400

    BUG/COMPAT: set tz on DatetimeIndex on pickle reconstruction (GH8367)

commit aa431bc166d9f936272eeee3b49d6b8f14c3945b
Merge: 6f4af67 3c1410f
Author: jreback <[email protected]>
Date:   Tue Sep 23 09:04:18 2014 -0400

    Merge pull request #8238 from stahlous/win_weights

    BUG: normalize rolling_window() weights, remove scikits-timeseries dependency for testing

commit 3c1410f56de973ea0e4932f37bb83461280b110e
Author: stahlous <[email protected]>
Date:   Sun Sep 7 12:59:56 2014 -0700

    BUG: rolling_window() properly averages weights in mean=True mode; removed scikits-timeseries dependency for testing; added further tests for rolling_window()

commit 6f4af67c1928c1b7336afbec3b627b23c6727bb4
Merge: c8e88d4 fcd1a94
Author: jreback <[email protected]>
Date:   Mon Sep 22 21:05:32 2014 -0400

    Merge pull request #8364 from jreback/datetime_ops

    BUG: Bug in alignment with TimeOps and non-unique indexes (GH8363)

commit fcd1a94a670b55394109553e7c6ca046144c813f
Author: jreback <[email protected]>
Date:   Mon Sep 22 18:06:51 2014 -0400

    BUG: Bug in alignment with TimeOps and non-unique indexes (GH8363)

commit c8e88d41acbb34ceec31d8dbf657623d5bcb6bbb
Merge: 709d0d6 998b8c9
Author: Tom Augspurger <[email protected]>
Date:   Mon Sep 22 16:46:12 2014 -0500

    Merge pull request #8353 from TomAugspurger/trellis-doc

    DOC: move trellis under visualization.rst

commit 709d0d6da18815f75a0592eb2cf3ee64129dc598
Author: jreback <[email protected]>
Date:   Mon Sep 22 15:33:36 2014 -0400

    DOC: timedelta resample docs

commit 2b9ecbe84df9919218488e5d2516e1a7a69505f6
Merge: 8069f47 d3f607d
Author: jreback <[email protected]>
Date:   Mon Sep 22 15:29:24 2014 -0400

    Merge pull request #8345 from jreback/resample

    ENH/BUG: allow timedelta resamples

commit d3f607da23e01abe96bcf2409914130bf80e4647
Author: jreback <[email protected]>
Date:   Sun Sep 21 18:25:21 2014 -0400

    ENH/BUG: allow timedelta resamples

commit 8069f4779cd1b0fe84abc53c73d8e6d0a44b0674
Merge: 1d65bc8 66d6912
Author: jreback <[email protected]>
Date:   Mon Sep 22 12:39:26 2014 -0400

    Merge pull request #8356 from jreback/excel_fix

    BUG: Work-around openpyxl-2.1.0 NumberFormat removal

commit 66d6912edb57b1509d428985f81437244a2e98c2
Author: John W. O'Brien <[email protected]>
Date:   Mon Sep 22 10:39:09 2014 -0400

    BUG: Work-around openpyxl-2.1.0 NumberFormat removal

commit 998b8c91c4fc57873508f6b7f5eb6a7107e32d86
Author: TomAugspurger <[email protected]>
Date:   Mon Sep 22 09:52:31 2014 -0500

    DOC: move trellis under visualization.rst

commit 1d65bc89d64c71f8d36f3ca92dd57db2efad7fdb
Merge: 0005e0f 3a98bc3
Author: Joris Van den Bossche <[email protected]>
Date:   Sun Sep 21 19:25:41 2014 +0200

    Merge pull request #8340 from jorisvandenbossche/sql-pep8

    CLN: pep8 clean up of sql.py

commit 0005e0f78e6086b544257648c96e476708d7c3e1
Author: jreback <[email protected]>
Date:   Sun Sep 21 13:21:18 2014 -0400

    TST:  test_precision_conversion on windows fix (GH8324)

commit 3a98bc381a05dfe8b2508b1612f8ccb9c0f779df
Author: Joris Van den Bossche <[email protected]>
Date:   Sun Sep 21 17:57:47 2014 +0200

    CLN: pep8 clean up of sql.py

commit 91e6dda3672d6fdb22d00e87d798ccc58ecd94f5
Merge: 7c319fd 7a1ffc7
Author: jreback <[email protected]>
Date:   Sat Sep 20 12:50:06 2014 -0400

    Merge pull request #8331 from jreback/concat_copy

    PERF: add copy=True argument to pd.concat to enable pass-thru concats with complete blocks (GH8252)

commit 7a1ffc74c54dec08c262b8996e35171fd41148e3
Author: jreback <[email protected]>
Date:   Sat Sep 20 11:59:36 2014 -0400

    PERF: add copy=True argument to pd.concat to enable pass-thru concats with complete blocks (GH8252)

commit 7c319fdb8fe8dce65683da5fe8418bdc3cf71f3e
Author: jreback <[email protected]>
Date:   Sat Sep 20 09:12:06 2014 -0400

    DOC: v0.15.0.txt edits

commit 034ef8dd4cef7862ffe1cb792b4244eab38fbd08
Merge: feb6870 99d8835
Author: jreback <[email protected]>
Date:   Sat Sep 20 04:53:31 2014 -0400

    Merge pull request #7565 from neirbowj/openpyxl2_issue7177

    ENH: Implement _Openpyxl2Writer for pandas.io.excel

commit feb68708eb395b4c55a905a080f522e1715b5a63
Author: jreback <[email protected]>
Date:   Fri Sep 19 19:27:31 2014 -0400

    DOC: v0.15.0.txt edits

commit 69fe8b3a70e15420cbd26896e6315fe48f4e36e0
Merge: 362db19 015e4d9
Author: jreback <[email protected]>
Date:   Fri Sep 19 16:05:54 2014 -0400

    Merge pull request #8322 from jreback/dtype

    BUG: Bug in casting when setting a column in a same-dtype block (GH7704)

commit 362db19249babd7b2169884bb7e3d43844776c13
Author: jreback <[email protected]>
Date:   Fri Sep 19 15:02:30 2014 -0400

    PERF: perf fixes

commit 015e4d9b532d2ff08f941647b1d15c03c7879353
Author: jreback <[email protected]>
Date:   Fri Sep 19 13:03:55 2014 -0400

    BUG: Bug in casting when setting a column in a same-dtype block (GH7704)

commit 744daa7278cf34bfcbf5a40e4f9153342370e0bc
Author: jreback <[email protected]>
Date:   Fri Sep 19 13:59:38 2014 -0400

    DOC: v0.15.0.txt corrections

commit 31c2558b33d727d6bd2fcb21f97e49d5a016513f
Author: Michael Mueller <[email protected]>
Date:   Fri Sep 19 11:36:22 2014 -0400

    Squashed commit of the following:

    commit 0e9d792fc9d5159179efd810a1092671dbbef3b1
    Author: Michael Mueller <[email protected]>
    Date:   Wed Sep 17 14:49:31 2014 -0400

        Added warnings about API changes

    commit 06472c21000b489841cc8e486ceddf05fd87a1c5
    Author: Michael Mueller <[email protected]>
    Date:   Fri Sep 12 22:36:06 2014 -0400

        Changed parameter name to skip_blank_lines

    commit afd3be30b4afcae0d9bc6278237aab6a4c9e7eb8
    Author: Michael Mueller <[email protected]>
    Date:   Fri Sep 12 21:50:08 2014 -0400

        Minor doc changes

    commit b47876e074f5f683a9a51768e480e24d9d3249ab
    Author: Michael Mueller <[email protected]>
    Date:   Fri Sep 12 19:26:22 2014 -0400

        Extended blank line skipping to custom line terminated/whitespace delimited reading

    commit 3f4a20a831b1bc0ca29779b315dc72d78ad2301e
    Author: Michael Mueller <[email protected]>
    Date:   Fri Sep 12 11:35:17 2014 -0400

        Changed around io docs section

    commit 223e17ecdcbe377cc69fd962221e03412f5e54d3
    Author: Michael Mueller <[email protected]>
    Date:   Tue Sep 9 23:13:37 2014 -0400

        Turned empty line skipping into a keyword parameter feature

    commit dcd31ca6bd0849eab87ea1c3c5441c8630ca3a35
    Author: Michael Mueller <[email protected]>
    Date:   Wed Sep 3 21:35:09 2014 -0400

        Squashed commit of the following:

        commit 9aea77954681c2f7d1336d94366221222d186c2b
        Author: Michael Mueller <[email protected]>
        Date:   Tue Aug 26 22:43:21 2014 -0400

            Fixed header/skiprows combination issue

        commit 1975affea3bf0bd6f1769a79e4b0c7fde17962df
        Author: Michael Mueller <[email protected]>
        Date:   Wed Jun 25 19:35:24 2014 -0400

            Added warning/notes about functionality change in docs, removed HTML changes

        commit 693c820092d9f17f9101074d29c2d7d53fa5a8ae
        Author: Michael Mueller <[email protected]>
        Date:   Wed Jun 25 15:38:41 2014 -0400

            Fixed problem with HTML reading and infinite loop in PythonParser __init__

        commit 2a0a4babac7a5e53279e…
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
Categorical Categorical Data Type
Projects
None yet
Milestone
0.15.0
3 participants
@jankatins @jorisvandenbossche @jreback