From jorisvandenbossche at gmail.com  Fri Apr 17 12:07:41 2015
From: jorisvandenbossche at gmail.com (Joris Van den Bossche)
Date: Fri, 17 Apr 2015 12:07:41 +0200
Subject: [Pandas-dev] Upcoming Index repr changes
Message-ID: <CALQtMBaj320BaSMfh4N6YeUJqP9s6=y7jSnwzJnFAwq36d=YsA@mail.gmail.com>

Hi all,

We have a PR pending to unify the string representation of the different
Index objects: https://github.com/pydata/pandas/pull/9901

What are the most important changes:

   - We propose to reduce the default number of values shown from 100 to 10
   (an option controllable as pd.options.display.max_seq_items).
   - The datetime-like indices (DatetimeIndex, TimedeltaIndex, PeriodIndex)
   were always somewhat different and get a new repr that is now more
   consistent with how it is for other Index types like Int64Index. This is
   the biggest change.

So for eg Int64Index not much changes (only 'name' is now also shown, and
the number of shown values has changed), but for DatetimeIndex the change
is larger.

*But we would like to get some feedback on this!*

Do you like the changes? For DatetimeIndex? For the number of shown values?
Would you want different behaviour for repr() and str()?

Some examples of the changes with the current state of the PR are shown
below:

Previous Behavior

In [1]: pd.get_option('max_seq_items')
Out[1]: 100

In [2]: pd.Index(range(4), name='foo')
Out[2]: Int64Index([0, 1, 2, 3], dtype='int64')

In [3]: pd.Index(range(104), name='foo')
Out[3]: Int64Index([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,
16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34,
35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53,
54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72,
73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91,
92, 93, 94, 95, 96, 97, 98, 99, ...], dtype='int64')

In [4]: pd.date_range('20130101', periods=4, name='foo', tz='US/Eastern')
Out[4]:
<class 'pandas.tseries.index.DatetimeIndex'>
[2013-01-01 00:00:00-05:00, ..., 2013-01-04 00:00:00-05:00]
Length: 4, Freq: D, Timezone: US/Eastern

In [5]: pd.date_range('20130101', periods=104, name='foo', tz='US/Eastern')
Out[5]:
<class 'pandas.tseries.index.DatetimeIndex'>
[2013-01-01 00:00:00-05:00, ..., 2013-04-14 00:00:00-04:00]
Length: 104, Freq: D, Timezone: US/Eastern

New Behavior

In [1]: pd.get_option('max_seq_items')
Out[1]: 10

In [9]: pd.Index(range(4), name='foo')
Out[9]: Int64Index([0, 1, 2, 3], dtype='int64', name=u'foo')

In [10]: pd.Index(range(104), name='foo')
Out[10]: Int64Index([0, 1, ..., 102, 103], dtype='int64', name=u'foo',
length=104)

In [11]: pd.date_range('20130101', periods=4, name='foo', tz='US/Eastern')
Out[11]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02
00:00:00-05:00', '2013-01-03 00:00:00-05:00', '2013-01-04 00:00:00-05:00'],
dtype='datetime64[ns]', name=u'foo', freq='D', tz='US/Eastern')

In [12]: pd.date_range('20130101', periods=104 ,name='foo', tz='US/Eastern')
Out[12]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02
00:00:00-05:00', ..., '2013-04-13 00:00:00-04:00', '2013-04-14
00:00:00-04:00'], dtype='datetime64[ns]', name=u'foo', length=104,
freq='D', tz='US/Eastern')
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pandas-dev/attachments/20150417/81ddb6ff/attachment.html>

From jeffreback at gmail.com  Fri Apr 17 19:09:18 2015
From: jeffreback at gmail.com (Jeff Reback)
Date: Fri, 17 Apr 2015 13:09:18 -0400
Subject: [Pandas-dev] sinhrks (Masaaki Horikoshi)
Message-ID: <CAHMnJKi4g+JcAHtkw4Jggu0ro2dSumh-N=ng+hp3VpG-zinDVg@mail.gmail.com>

All,

@sinhrks (Masaaki Horikoshi)

has accepted the invitation to be a *pandas core dev!*

Thanks for all of the effort you have put in.

Let keep it up to get even more people into pandas contributions.

This project is quite popular, from just seeing the reaction at PyCon and
is used in a lot

of different fields / ways.

Jeff
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pandas-dev/attachments/20150417/16156e05/attachment.html>

From wesmckinn at gmail.com  Sat Apr 18 00:27:44 2015
From: wesmckinn at gmail.com (Wes McKinney)
Date: Fri, 17 Apr 2015 15:27:44 -0700
Subject: [Pandas-dev] sinhrks (Masaaki Horikoshi)
In-Reply-To: <CAHMnJKi4g+JcAHtkw4Jggu0ro2dSumh-N=ng+hp3VpG-zinDVg@mail.gmail.com>
References: <CAHMnJKi4g+JcAHtkw4Jggu0ro2dSumh-N=ng+hp3VpG-zinDVg@mail.gmail.com>
Message-ID: <CAJPUwMD_3-6g0kOYhmeAR4Ls=r_be1VZzS4B9kjBQzFDv42hZA@mail.gmail.com>

Great news -- @sinhrks, welcome to the team, and thanks for all the
hard work on the project!

cheers,
Wes

On Fri, Apr 17, 2015 at 10:09 AM, Jeff Reback <jeffreback at gmail.com> wrote:
> All,
>
> @sinhrks (Masaaki Horikoshi)
>
> has accepted the invitation to be a pandas core dev!
>
> Thanks for all of the effort you have put in.
>
> Let keep it up to get even more people into pandas contributions.
>
> This project is quite popular, from just seeing the reaction at PyCon and is
> used in a lot
>
> of different fields / ways.
>
> Jeff
>
> _______________________________________________
> Pandas-dev mailing list
> Pandas-dev at python.org
> https://mail.python.org/mailman/listinfo/pandas-dev
>

From shoyer at gmail.com  Sat Apr 18 00:33:59 2015
From: shoyer at gmail.com (Stephan Hoyer)
Date: Fri, 17 Apr 2015 15:33:59 -0700
Subject: [Pandas-dev] sinhrks (Masaaki Horikoshi)
In-Reply-To: <CAJPUwMD_3-6g0kOYhmeAR4Ls=r_be1VZzS4B9kjBQzFDv42hZA@mail.gmail.com>
References: <CAHMnJKi4g+JcAHtkw4Jggu0ro2dSumh-N=ng+hp3VpG-zinDVg@mail.gmail.com>
 <CAJPUwMD_3-6g0kOYhmeAR4Ls=r_be1VZzS4B9kjBQzFDv42hZA@mail.gmail.com>
Message-ID: <CAEQ_TvfgBADxRr6b3BcgCvFaEKawBFRTDRFBmsMK86b6sLCHsA@mail.gmail.com>

Congratulations Masaaki! We're really happy to have you on board :).

On Fri, Apr 17, 2015 at 3:27 PM, Wes McKinney <wesmckinn at gmail.com> wrote:

> Great news -- @sinhrks, welcome to the team, and thanks for all the
> hard work on the project!
>
> cheers,
> Wes
>
> On Fri, Apr 17, 2015 at 10:09 AM, Jeff Reback <jeffreback at gmail.com>
> wrote:
> > All,
> >
> > @sinhrks (Masaaki Horikoshi)
> >
> > has accepted the invitation to be a pandas core dev!
> >
> > Thanks for all of the effort you have put in.
> >
> > Let keep it up to get even more people into pandas contributions.
> >
> > This project is quite popular, from just seeing the reaction at PyCon
> and is
> > used in a lot
> >
> > of different fields / ways.
> >
> > Jeff
> >
> > _______________________________________________
> > Pandas-dev mailing list
> > Pandas-dev at python.org
> > https://mail.python.org/mailman/listinfo/pandas-dev
> >
> _______________________________________________
> Pandas-dev mailing list
> Pandas-dev at python.org
> https://mail.python.org/mailman/listinfo/pandas-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pandas-dev/attachments/20150417/435a9f0d/attachment.html>

From jorisvandenbossche at gmail.com  Tue Apr 21 02:59:33 2015
From: jorisvandenbossche at gmail.com (Joris Van den Bossche)
Date: Tue, 21 Apr 2015 02:59:33 +0200
Subject: [Pandas-dev] [pydata] Re: Upcoming Index repr changes
In-Reply-To: <6ab996ef-e112-4f52-9d58-60947629011f@googlegroups.com>
References: <CALQtMBaj320BaSMfh4N6YeUJqP9s6=y7jSnwzJnFAwq36d=YsA@mail.gmail.com>
 <3cb616db-fbad-44a9-970d-93c7dda3d42d@googlegroups.com>
 <6ab996ef-e112-4f52-9d58-60947629011f@googlegroups.com>
Message-ID: <CALQtMBYgJumi=Jv6DFjOOyFnkvfcU1ffrk=M_pN4m_hsepWEWw@mail.gmail.com>

I like the suggestion of John to have something more like the output of
numpy arrays.

For example, the proposed repr:

In [12]: pd.date_range('20130101',periods=104,name='foo',tz='US/Eastern')
Out[12]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02
00:00:00-05:00', ..., '2013-04-13 00:00:00-04:00', '2013-04-14
00:00:00-04:00'], dtype='datetime64[ns]', name=u'foo', length=104,
freq='D', tz='US/Eastern')

would then be something like this:

In [12]: pd.date_range('20130101',periods=104,name='foo',tz='US/Eastern')
Out[12]:
DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 00:00:00-05:00',
...,
               '2013-04-13 00:00:00-04:00', '2013-04-14 00:00:00-04:00'],
              dtype='datetime64[ns]', name=u'foo', length=104, freq='D',
tz='US/Eastern')


2015-04-21 2:53 GMT+02:00 Jeff <jeffreback at gmail.com>:

>
> John, you are quoting the current impl (which is first), the new is like
> this:
>
> In [11]: pd.date_range('20130101',periods=4,name='foo',tz='US/Eastern')
> Out[11]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 00:00:00-05:00', '2013-01-03 00:00:00-05:00', '2013-01-04 00:00:00-05:00'], dtype='datetime64[ns]', name=u'foo', freq='D', tz='US/Eastern')
>
> In [12]: pd.date_range('20130101',periods=104,name='foo',tz='US/Eastern')
> Out[12]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 00:00:00-05:00', ..., '2013-04-13 00:00:00-04:00', '2013-04-14 00:00:00-04:00'], dtype='datetime64[ns]', name=u'foo', length=104, freq='D', tz='US/Eastern')
>
> Lorenzo, to answer your question, MultiIndexes are unchanged (and
> CategoricalIndex are new). We *could* make them a single line but would be
> pretty crowded.
>
> Note that MultiIndex and CategoricalIndex are multi-line repr and do no
> truncate sequences (of e.g. labels), this is consistent with previous
> versions. (easy to change this though)
>
> In [1]: MultiIndex.from_product([list('abcdefg'),range(10)],names=['first','second'])
> Out[1]:
> MultiIndex(levels=[[u'a', u'b', u'c', u'd', u'e', u'f', u'g'], [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]],
>            labels=[[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6], [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9]],
>            names=[u'first', u'second'])
>
> In [4]: pd.CategoricalIndex(np.random.randint(0,5,size=100),name='foo')
> Out[4]:
> CategoricalIndex([3, 0, 0, 3, 1, 3, 0, 4, 2, 3, 0, 4, 0, 1, 2, 0, 4, 1, 4, 2, 3, 1, 0, 4, 4, 3, 0, 3, 0, 1, 2, 3, 3, 1, 1, 0, 0, 4, 4, 1, 1, 3, 1, 1, 4, 4, 3, 0, 0, 0, 4, 4, 0, 1, 3, 1, 2, 0, 3, 1, 2, 2, 2, 1, 1, 4, 1, 0, 4, 3, 3, 0, 0, 0, 4, 4, 1, 4, 2, 2, 1, 4, 0, 0, 0, 4, 3, 0, 4, 0, 0, 0, 3, 3, 1, 2, 2, 3, 4, 1],
>                  categories=[0, 1, 2, 3, 4],
>                  ordered=False,
>                  name=u'foo',
>                  dtype='category')
>
>
>
>
>
> On Monday, April 20, 2015 at 8:37:01 PM UTC-4, John E wrote:
>>
>> This is probably not the sort of comment you're looking for, but I'd like
>> to see more of a table-style output.  I can just put a 'values' at the end
>> to get the more numpy like output (which is easier to read IMO), but it
>> won't stop at 10 or 100 unless I tell it to.  Nevertheless, I think it's
>> much easer to read this:
>>
>> pd.date_range('20130101', periods=104, name='foo', tz='US/Eastern').values
>> Out[442]:
>> array(['2013-01-01T00:00:00.000000000-0500',
>>        '2013-01-02T00:00:00.000000000-0500',
>>        '2013-01-03T00:00:00.000000000-0500',
>>        '2013-01-04T00:00:00.000000000-0500',
>>        '2013-01-05T00:00:00.000000000-0500',
>>
>> than this:
>>
>> pd.date_range('20130101', periods=104, name='foo', tz='US/Eastern')
>> Out[443]:
>> <class 'pandas.tseries.index.DatetimeIndex'>
>> [2013-01-01 00:00:00-05:00, ..., 2013-04-14 00:00:00-04:00]
>> Length: 104, Freq: D, Timezone: US/Eastern
>>
>>
>> On Friday, April 17, 2015 at 6:07:44 AM UTC-4, Joris Van den Bossche
>> wrote:
>>>
>>> Hi all,
>>>
>>> We have a PR pending to unify the string representation of the different
>>> Index objects: https://github.com/pydata/pandas/pull/9901
>>>
>>> What are the most important changes:
>>>
>>>    - We propose to reduce the default number of values shown from 100
>>>    to 10 (an option controllable as pd.options.display.max_seq_items).
>>>    - The datetime-like indices (DatetimeIndex, TimedeltaIndex,
>>>    PeriodIndex) were always somewhat different and get a new repr that is now
>>>    more consistent with how it is for other Index types like Int64Index. This
>>>    is the biggest change.
>>>
>>> So for eg Int64Index not much changes (only 'name' is now also shown,
>>> and the number of shown values has changed), but for DatetimeIndex the
>>> change is larger.
>>>
>>> *But we would like to get some feedback on this!*
>>>
>>> Do you like the changes? For DatetimeIndex? For the number of shown
>>> values?
>>> Would you want different behaviour for repr() and str()?
>>>
>>> Some examples of the changes with the current state of the PR are shown
>>> below:
>>>
>>> Previous Behavior
>>>
>>> In [1]: pd.get_option('max_seq_items')
>>> Out[1]: 100
>>>
>>> In [2]: pd.Index(range(4), name='foo')
>>> Out[2]: Int64Index([0, 1, 2, 3], dtype='int64')
>>>
>>> In [3]: pd.Index(range(104), name='foo')
>>> Out[3]: Int64Index([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,
>>> 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33,
>>> 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52,
>>> 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71,
>>> 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90,
>>> 91, 92, 93, 94, 95, 96, 97, 98, 99, ...], dtype='int64')
>>>
>>> In [4]: pd.date_range('20130101', periods=4, name='foo', tz='US/Eastern')
>>> Out[4]:
>>> <class 'pandas.tseries.index.DatetimeIndex'>
>>> [2013-01-01 00:00:00-05:00, ..., 2013-01-04 00:00:00-05:00]
>>> Length: 4, Freq: D, Timezone: US/Eastern
>>>
>>> In [5]: pd.date_range('20130101', periods=104, name='foo',
>>> tz='US/Eastern')
>>> Out[5]:
>>> <class 'pandas.tseries.index.DatetimeIndex'>
>>> [2013-01-01 00:00:00-05:00, ..., 2013-04-14 00:00:00-04:00]
>>> Length: 104, Freq: D, Timezone: US/Eastern
>>>
>>> New Behavior
>>>
>>> In [1]: pd.get_option('max_seq_items')
>>> Out[1]: 10
>>>
>>> In [9]: pd.Index(range(4), name='foo')
>>> Out[9]: Int64Index([0, 1, 2, 3], dtype='int64', name=u'foo')
>>>
>>> In [10]: pd.Index(range(104), name='foo')
>>> Out[10]: Int64Index([0, 1, ..., 102, 103], dtype='int64', name=u'foo',
>>> length=104)
>>>
>>> In [11]: pd.date_range('20130101', periods=4, name='foo',
>>> tz='US/Eastern')
>>> Out[11]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02
>>> 00:00:00-05:00', '2013-01-03 00:00:00-05:00', '2013-01-04 00:00:00-05:00'],
>>> dtype='datetime64[ns]', name=u'foo', freq='D', tz='US/Eastern')
>>>
>>> In [12]: pd.date_range('20130101', periods=104 ,name='foo',
>>> tz='US/Eastern')
>>> Out[12]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02
>>> 00:00:00-05:00', ..., '2013-04-13 00:00:00-04:00', '2013-04-14
>>> 00:00:00-04:00'], dtype='datetime64[ns]', name=u'foo', length=104,
>>> freq='D', tz='US/Eastern')
>>>
>>>  --
> You received this message because you are subscribed to the Google Groups
> "PyData" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to pydata+unsubscribe at googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pandas-dev/attachments/20150421/78b2ec5f/attachment.html>

From jeffreback at gmail.com  Tue Apr 21 02:53:17 2015
From: jeffreback at gmail.com (Jeff)
Date: Mon, 20 Apr 2015 17:53:17 -0700 (PDT)
Subject: [Pandas-dev] Upcoming Index repr changes
In-Reply-To: <3cb616db-fbad-44a9-970d-93c7dda3d42d@googlegroups.com>
References: <CALQtMBaj320BaSMfh4N6YeUJqP9s6=y7jSnwzJnFAwq36d=YsA@mail.gmail.com>
 <3cb616db-fbad-44a9-970d-93c7dda3d42d@googlegroups.com>
Message-ID: <6ab996ef-e112-4f52-9d58-60947629011f@googlegroups.com>


John, you are quoting the current impl (which is first), the new is like 
this:

In [11]: pd.date_range('20130101',periods=4,name='foo',tz='US/Eastern')
Out[11]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 00:00:00-05:00', '2013-01-03 00:00:00-05:00', '2013-01-04 00:00:00-05:00'], dtype='datetime64[ns]', name=u'foo', freq='D', tz='US/Eastern')

In [12]: pd.date_range('20130101',periods=104,name='foo',tz='US/Eastern')
Out[12]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 00:00:00-05:00', ..., '2013-04-13 00:00:00-04:00', '2013-04-14 00:00:00-04:00'], dtype='datetime64[ns]', name=u'foo', length=104, freq='D', tz='US/Eastern')

Lorenzo, to answer your question, MultiIndexes are unchanged (and 
CategoricalIndex are new). We *could* make them a single line but would be 
pretty crowded. 

Note that MultiIndex and CategoricalIndex are multi-line repr and do no 
truncate sequences (of e.g. labels), this is consistent with previous 
versions. (easy to change this though)

In [1]: MultiIndex.from_product([list('abcdefg'),range(10)],names=['first','second'])
Out[1]: 
MultiIndex(levels=[[u'a', u'b', u'c', u'd', u'e', u'f', u'g'], [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]],
           labels=[[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6], [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9]],
           names=[u'first', u'second'])

In [4]: pd.CategoricalIndex(np.random.randint(0,5,size=100),name='foo')
Out[4]: 
CategoricalIndex([3, 0, 0, 3, 1, 3, 0, 4, 2, 3, 0, 4, 0, 1, 2, 0, 4, 1, 4, 2, 3, 1, 0, 4, 4, 3, 0, 3, 0, 1, 2, 3, 3, 1, 1, 0, 0, 4, 4, 1, 1, 3, 1, 1, 4, 4, 3, 0, 0, 0, 4, 4, 0, 1, 3, 1, 2, 0, 3, 1, 2, 2, 2, 1, 1, 4, 1, 0, 4, 3, 3, 0, 0, 0, 4, 4, 1, 4, 2, 2, 1, 4, 0, 0, 0, 4, 3, 0, 4, 0, 0, 0, 3, 3, 1, 2, 2, 3, 4, 1],
                 categories=[0, 1, 2, 3, 4],
                 ordered=False,
                 name=u'foo',
                 dtype='category')





On Monday, April 20, 2015 at 8:37:01 PM UTC-4, John E wrote:
>
> This is probably not the sort of comment you're looking for, but I'd like 
> to see more of a table-style output.  I can just put a 'values' at the end 
> to get the more numpy like output (which is easier to read IMO), but it 
> won't stop at 10 or 100 unless I tell it to.  Nevertheless, I think it's 
> much easer to read this:
>
> pd.date_range('20130101', periods=104, name='foo', tz='US/Eastern').values
> Out[442]: 
> array(['2013-01-01T00:00:00.000000000-0500',
>        '2013-01-02T00:00:00.000000000-0500',
>        '2013-01-03T00:00:00.000000000-0500',
>        '2013-01-04T00:00:00.000000000-0500',
>        '2013-01-05T00:00:00.000000000-0500',
>
> than this:
>
> pd.date_range('20130101', periods=104, name='foo', tz='US/Eastern')
> Out[443]: 
> <class 'pandas.tseries.index.DatetimeIndex'>
> [2013-01-01 00:00:00-05:00, ..., 2013-04-14 00:00:00-04:00]
> Length: 104, Freq: D, Timezone: US/Eastern
>
>
> On Friday, April 17, 2015 at 6:07:44 AM UTC-4, Joris Van den Bossche wrote:
>>
>> Hi all,
>>
>> We have a PR pending to unify the string representation of the different 
>> Index objects: https://github.com/pydata/pandas/pull/9901
>>
>> What are the most important changes:
>>
>>    - We propose to reduce the default number of values shown from 100 to 
>>    10 (an option controllable as pd.options.display.max_seq_items). 
>>    - The datetime-like indices (DatetimeIndex, TimedeltaIndex, 
>>    PeriodIndex) were always somewhat different and get a new repr that is now 
>>    more consistent with how it is for other Index types like Int64Index. This 
>>    is the biggest change.
>>
>> So for eg Int64Index not much changes (only 'name' is now also shown, and 
>> the number of shown values has changed), but for DatetimeIndex the change 
>> is larger.
>>
>> *But we would like to get some feedback on this!*
>>
>> Do you like the changes? For DatetimeIndex? For the number of shown 
>> values?
>> Would you want different behaviour for repr() and str()?
>>
>> Some examples of the changes with the current state of the PR are shown 
>> below:
>>
>> Previous Behavior
>>
>> In [1]: pd.get_option('max_seq_items')
>> Out[1]: 100
>>
>> In [2]: pd.Index(range(4), name='foo')
>> Out[2]: Int64Index([0, 1, 2, 3], dtype='int64')
>>
>> In [3]: pd.Index(range(104), name='foo')
>> Out[3]: Int64Index([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 
>> 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 
>> 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 
>> 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 
>> 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 
>> 92, 93, 94, 95, 96, 97, 98, 99, ...], dtype='int64')
>>
>> In [4]: pd.date_range('20130101', periods=4, name='foo', tz='US/Eastern')
>> Out[4]:
>> <class 'pandas.tseries.index.DatetimeIndex'>
>> [2013-01-01 00:00:00-05:00, ..., 2013-01-04 00:00:00-05:00]
>> Length: 4, Freq: D, Timezone: US/Eastern
>>
>> In [5]: pd.date_range('20130101', periods=104, name='foo', 
>> tz='US/Eastern')
>> Out[5]:
>> <class 'pandas.tseries.index.DatetimeIndex'>
>> [2013-01-01 00:00:00-05:00, ..., 2013-04-14 00:00:00-04:00]
>> Length: 104, Freq: D, Timezone: US/Eastern
>>
>> New Behavior
>>
>> In [1]: pd.get_option('max_seq_items')
>> Out[1]: 10
>>
>> In [9]: pd.Index(range(4), name='foo')
>> Out[9]: Int64Index([0, 1, 2, 3], dtype='int64', name=u'foo')
>>
>> In [10]: pd.Index(range(104), name='foo')
>> Out[10]: Int64Index([0, 1, ..., 102, 103], dtype='int64', name=u'foo', 
>> length=104)
>>
>> In [11]: pd.date_range('20130101', periods=4, name='foo', tz='US/Eastern')
>> Out[11]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 
>> 00:00:00-05:00', '2013-01-03 00:00:00-05:00', '2013-01-04 00:00:00-05:00'], 
>> dtype='datetime64[ns]', name=u'foo', freq='D', tz='US/Eastern')
>>
>> In [12]: pd.date_range('20130101', periods=104 ,name='foo', 
>> tz='US/Eastern')
>> Out[12]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 
>> 00:00:00-05:00', ..., '2013-04-13 00:00:00-04:00', '2013-04-14 
>> 00:00:00-04:00'], dtype='datetime64[ns]', name=u'foo', length=104, 
>> freq='D', tz='US/Eastern')
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pandas-dev/attachments/20150420/7e29f1aa/attachment-0001.html>

From eiler13 at gmail.com  Mon Apr 20 20:37:09 2015
From: eiler13 at gmail.com (John E)
Date: Tue, 21 Apr 2015 00:37:09 -0000
Subject: [Pandas-dev] Upcoming Index repr changes
In-Reply-To: <CALQtMBaj320BaSMfh4N6YeUJqP9s6=y7jSnwzJnFAwq36d=YsA@mail.gmail.com>
References: <CALQtMBaj320BaSMfh4N6YeUJqP9s6=y7jSnwzJnFAwq36d=YsA@mail.gmail.com>
Message-ID: <3cb616db-fbad-44a9-970d-93c7dda3d42d@googlegroups.com>

This is probably not the sort of comment you're looking for, but I'd like 
to see more of a table-style output.  I can just put a 'values' at the end 
to get the more numpy like output (which is easier to read IMO), but it 
won't stop at 10 or 100 unless I tell it to.  Nevertheless, I think it's 
much easer to read this:

pd.date_range('20130101', periods=104, name='foo', tz='US/Eastern').values
Out[442]: 
array(['2013-01-01T00:00:00.000000000-0500',
       '2013-01-02T00:00:00.000000000-0500',
       '2013-01-03T00:00:00.000000000-0500',
       '2013-01-04T00:00:00.000000000-0500',
       '2013-01-05T00:00:00.000000000-0500',

than this:

pd.date_range('20130101', periods=104, name='foo', tz='US/Eastern')
Out[443]: 
<class 'pandas.tseries.index.DatetimeIndex'>
[2013-01-01 00:00:00-05:00, ..., 2013-04-14 00:00:00-04:00]
Length: 104, Freq: D, Timezone: US/Eastern


On Friday, April 17, 2015 at 6:07:44 AM UTC-4, Joris Van den Bossche wrote:
>
> Hi all,
>
> We have a PR pending to unify the string representation of the different 
> Index objects: https://github.com/pydata/pandas/pull/9901
>
> What are the most important changes:
>
>    - We propose to reduce the default number of values shown from 100 to 
>    10 (an option controllable as pd.options.display.max_seq_items). 
>    - The datetime-like indices (DatetimeIndex, TimedeltaIndex, 
>    PeriodIndex) were always somewhat different and get a new repr that is now 
>    more consistent with how it is for other Index types like Int64Index. This 
>    is the biggest change.
>
> So for eg Int64Index not much changes (only 'name' is now also shown, and 
> the number of shown values has changed), but for DatetimeIndex the change 
> is larger.
>
> *But we would like to get some feedback on this!*
>
> Do you like the changes? For DatetimeIndex? For the number of shown values?
> Would you want different behaviour for repr() and str()?
>
> Some examples of the changes with the current state of the PR are shown 
> below:
>
> Previous Behavior
>
> In [1]: pd.get_option('max_seq_items')
> Out[1]: 100
>
> In [2]: pd.Index(range(4), name='foo')
> Out[2]: Int64Index([0, 1, 2, 3], dtype='int64')
>
> In [3]: pd.Index(range(104), name='foo')
> Out[3]: Int64Index([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 
> 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 
> 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 
> 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 
> 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 
> 92, 93, 94, 95, 96, 97, 98, 99, ...], dtype='int64')
>
> In [4]: pd.date_range('20130101', periods=4, name='foo', tz='US/Eastern')
> Out[4]:
> <class 'pandas.tseries.index.DatetimeIndex'>
> [2013-01-01 00:00:00-05:00, ..., 2013-01-04 00:00:00-05:00]
> Length: 4, Freq: D, Timezone: US/Eastern
>
> In [5]: pd.date_range('20130101', periods=104, name='foo', tz='US/Eastern')
> Out[5]:
> <class 'pandas.tseries.index.DatetimeIndex'>
> [2013-01-01 00:00:00-05:00, ..., 2013-04-14 00:00:00-04:00]
> Length: 104, Freq: D, Timezone: US/Eastern
>
> New Behavior
>
> In [1]: pd.get_option('max_seq_items')
> Out[1]: 10
>
> In [9]: pd.Index(range(4), name='foo')
> Out[9]: Int64Index([0, 1, 2, 3], dtype='int64', name=u'foo')
>
> In [10]: pd.Index(range(104), name='foo')
> Out[10]: Int64Index([0, 1, ..., 102, 103], dtype='int64', name=u'foo', 
> length=104)
>
> In [11]: pd.date_range('20130101', periods=4, name='foo', tz='US/Eastern')
> Out[11]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 
> 00:00:00-05:00', '2013-01-03 00:00:00-05:00', '2013-01-04 00:00:00-05:00'], 
> dtype='datetime64[ns]', name=u'foo', freq='D', tz='US/Eastern')
>
> In [12]: pd.date_range('20130101', periods=104 ,name='foo', 
> tz='US/Eastern')
> Out[12]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 
> 00:00:00-05:00', ..., '2013-04-13 00:00:00-04:00', '2013-04-14 
> 00:00:00-04:00'], dtype='datetime64[ns]', name=u'foo', length=104, 
> freq='D', tz='US/Eastern')
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pandas-dev/attachments/20150421/e555dea4/attachment.html>

From lorenzo.deleo at gmail.com  Mon Apr 20 17:13:33 2015
From: lorenzo.deleo at gmail.com (Lorenzo De Leo)
Date: Mon, 20 Apr 2015 21:13:33 -0000
Subject: [Pandas-dev] Upcoming Index repr changes
In-Reply-To: <CALQtMBaj320BaSMfh4N6YeUJqP9s6=y7jSnwzJnFAwq36d=YsA@mail.gmail.com>
References: <CALQtMBaj320BaSMfh4N6YeUJqP9s6=y7jSnwzJnFAwq36d=YsA@mail.gmail.com>
Message-ID: <db38c38a-8c74-4143-be8b-49b995b8f931@googlegroups.com>

I like the changes you propose, the new version is much more readable. I 
used to be wary of calling df.index because it can be slow and the output 
is a bit messy, and I'm usually too lazy to select just a slice of it, so 
having something like this done by default is a welcome change.
Just a question, does it apply also to multiindexes?

Cheers!


On Friday, April 17, 2015 at 12:07:44 PM UTC+2, Joris Van den Bossche wrote:
>
> Hi all,
>
> We have a PR pending to unify the string representation of the different 
> Index objects: https://github.com/pydata/pandas/pull/9901
>
> What are the most important changes:
>
>    - We propose to reduce the default number of values shown from 100 to 
>    10 (an option controllable as pd.options.display.max_seq_items). 
>    - The datetime-like indices (DatetimeIndex, TimedeltaIndex, 
>    PeriodIndex) were always somewhat different and get a new repr that is now 
>    more consistent with how it is for other Index types like Int64Index. This 
>    is the biggest change.
>
> So for eg Int64Index not much changes (only 'name' is now also shown, and 
> the number of shown values has changed), but for DatetimeIndex the change 
> is larger.
>
> *But we would like to get some feedback on this!*
>
> Do you like the changes? For DatetimeIndex? For the number of shown values?
> Would you want different behaviour for repr() and str()?
>
> Some examples of the changes with the current state of the PR are shown 
> below:
>
> Previous Behavior
>
> In [1]: pd.get_option('max_seq_items')
> Out[1]: 100
>
> In [2]: pd.Index(range(4), name='foo')
> Out[2]: Int64Index([0, 1, 2, 3], dtype='int64')
>
> In [3]: pd.Index(range(104), name='foo')
> Out[3]: Int64Index([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 
> 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 
> 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 
> 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 
> 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 
> 92, 93, 94, 95, 96, 97, 98, 99, ...], dtype='int64')
>
> In [4]: pd.date_range('20130101', periods=4, name='foo', tz='US/Eastern')
> Out[4]:
> <class 'pandas.tseries.index.DatetimeIndex'>
> [2013-01-01 00:00:00-05:00, ..., 2013-01-04 00:00:00-05:00]
> Length: 4, Freq: D, Timezone: US/Eastern
>
> In [5]: pd.date_range('20130101', periods=104, name='foo', tz='US/Eastern')
> Out[5]:
> <class 'pandas.tseries.index.DatetimeIndex'>
> [2013-01-01 00:00:00-05:00, ..., 2013-04-14 00:00:00-04:00]
> Length: 104, Freq: D, Timezone: US/Eastern
>
> New Behavior
>
> In [1]: pd.get_option('max_seq_items')
> Out[1]: 10
>
> In [9]: pd.Index(range(4), name='foo')
> Out[9]: Int64Index([0, 1, 2, 3], dtype='int64', name=u'foo')
>
> In [10]: pd.Index(range(104), name='foo')
> Out[10]: Int64Index([0, 1, ..., 102, 103], dtype='int64', name=u'foo', 
> length=104)
>
> In [11]: pd.date_range('20130101', periods=4, name='foo', tz='US/Eastern')
> Out[11]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 
> 00:00:00-05:00', '2013-01-03 00:00:00-05:00', '2013-01-04 00:00:00-05:00'], 
> dtype='datetime64[ns]', name=u'foo', freq='D', tz='US/Eastern')
>
> In [12]: pd.date_range('20130101', periods=104 ,name='foo', 
> tz='US/Eastern')
> Out[12]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 
> 00:00:00-05:00', ..., '2013-04-13 00:00:00-04:00', '2013-04-14 
> 00:00:00-04:00'], dtype='datetime64[ns]', name=u'foo', length=104, 
> freq='D', tz='US/Eastern')
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pandas-dev/attachments/20150420/34901d6e/attachment-0001.html>