[Pandas-dev] Upcoming Index repr changes

[Lorenzo De Leo]

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>