Finding if atleast one entry from a specified month string exists in Pandas datetime index

Question

I have a pandas dataframe test_df that looks like this:

                            reading                               
01-Jan-2016 00:00:00.000  20.464020
02-Jan-2016 00:00:00.000  22.440950
03-Jan-2016 00:00:00.000  27.181500
04-Jan-2016 00:00:00.000  25.318260
05-Jan-2016 00:00:00.000  25.376050
06-Jan-2016 00:00:00.000   0.067112
07-Jan-2016 00:00:00.000  19.313950
08-Jan-2016 00:00:00.000  26.677340
09-Jan-2016 00:00:00.000  26.801620
10-Jan-2016 00:00:00.000  22.583950
11-Jan-2016 00:00:00.000   0.002765
12-Jan-2016 00:00:00.000  26.496440
13-Jan-2016 00:00:00.000  23.233720
14-Jan-2016 00:00:00.000  23.956080
15-Jan-2016 00:00:00.000  26.958120
16-Jan-2016 00:00:00.000  27.351270
17-Jan-2016 00:00:00.000  28.348710
18-Jan-2016 00:00:00.000  25.494090
19-Jan-2016 00:00:00.000  26.342880
20-Jan-2016 00:00:00.000  24.645530

The problem: given a string like '2016-01' a.k.a 'yyyy-mm', I want to find out if any entry from the specified month is present in the index of the pandas dataframe test_df.

What I'm expecting is True for '2016-01' and False for any other string. Looking for the most concise method to do that.

Problem setup:

To make things easy, this is the code to obtain test dataframe:

import pandas as pd
temp_df = pd.read_json('{"reading":{"01-Jan-2016 00:00:00.000":20.46402,"02-Jan-2016 00:00:00.000":22.44095,"03-Jan-2016 00:00:00.000":27.1815,"04-Jan-2016 00:00:00.000":25.31826,"05-Jan-2016 00:00:00.000":25.37605,"06-Jan-2016 00:00:00.000":0.06711243,"07-Jan-2016 00:00:00.000":19.31395,"08-Jan-2016 00:00:00.000":26.67734,"09-Jan-2016 00:00:00.000":26.80162,"10-Jan-2016 00:00:00.000":22.58395,"11-Jan-2016 00:00:00.000":0.002765084,"12-Jan-2016 00:00:00.000":26.49644,"13-Jan-2016 00:00:00.000":23.23372,"14-Jan-2016 00:00:00.000":23.95608,"15-Jan-2016 00:00:00.000":26.95812,"16-Jan-2016 00:00:00.000":27.35127,"17-Jan-2016 00:00:00.000":28.34871,"18-Jan-2016 00:00:00.000":25.49409,"19-Jan-2016 00:00:00.000":26.34288,"20-Jan-2016 00:00:00.000":24.64553}}')

I've tried:

>>'2016-01' in test_df.index
False

score 2 · Accepted Answer · edited May 23 '17 at 10:30

2

If use DatetimeIndex you can use to_period for convert to PeriodIndex and then any (thank you John Zwinck ):

print (temp_df.index.to_period('m'))
PeriodIndex(['2016-01', '2016-01', '2016-01', '2016-01', '2016-01', '2016-01',
             '2016-01', '2016-01', '2016-01', '2016-01', '2016-01', '2016-01',
             '2016-01', '2016-01', '2016-01', '2016-01', '2016-01', '2016-01',
             '2016-01', '2016-01'],
            dtype='period[M]', freq='M')

print (temp_df.index.to_period('m') == '2016-01')
[ True  True  True  True  True  True  True  True  True  True  True  True
  True  True  True  True  True  True  True  True]

print ((temp_df.index.to_period('m') == '2016-01').any())
True

edited May 23 '17 at 10:30

Community

1
1

answered Dec 01 '16 at 11:53

jezrael

822,522
95
1,334
1,252

2

`to_period` is unnecessary here, `DatetimeIndex` supports partial datetime string matching see the [docs](http://pandas.pydata.org/pandas-docs/stable/timeseries.html#datetimeindex-partial-string-indexing) – EdChum Dec 01 '16 at 11:55
But why different output? Why you get no all True? – jezrael Dec 01 '16 at 11:57
It seems no support it. – jezrael Dec 01 '16 at 11:58
1

Hmm, maybe this is a bug will investigate – EdChum Dec 01 '16 at 12:00
It ssems convert `2016-01` to day and year, what is wrong. – jezrael Dec 01 '16 at 12:00
2

Simpler is `temp_df.index.to_period('m') == '2016-01'` which also works. OP didn't say there was a list of periods to look for, just one. – John Zwinck Dec 01 '16 at 12:03
3

It looks like if you pass even `2016` to `isin` it's up-converting this to a full datetime object with default attributes so it becomes `2016-01-01` so only the first entry is `True`, it's unclear if this is expected or not – EdChum Dec 01 '16 at 12:04
Even `temp_df.index.strftime('%Y-%m') == '2016-01'` FYI just does the job. – Nickil Maveli Dec 01 '16 at 12:05
@NickilMaveli: That's not efficient. – John Zwinck Dec 01 '16 at 12:06

Finding if atleast one entry from a specified month string exists in Pandas datetime index

Problem setup:

1 Answers1