CODEWITHSUNDEEP
pythonpandas
Zanam
Zanam

Reputation: 4807

pandas replacing specific elements from a tseries datetimeindex

I have pandas object as:

df_days
DatetimeIndex(['2015-05-24', '2015-05-24', '2015-05-24',..., '2016-03-19',
               '2016-03-19', '2016-03-20'],
              dtype='datetime64[ns]', length=7224, freq=None)

type(df_days)
<class 'pandas.tseries.index.DatetimeIndex'>

I have an array of indexes as:

hrIdx
(array([  23,   47,   71,   95,  119,  143,  167,  191,  215,  239,  263,
        287,  311,  335,  359,  383,  407,  431,  455,  479,  503,  527,
        551,  575,  599, ...,1592], dtype=int64),)

I am trying the following:

df_days[hrIdx] = df_days[hrIdx] + td(days=-1)

But the assignment operation fails. The right side operation of substracting a day works fine.

Error is:

    raise TypeError("Index does not support mutable operations")
TypeError: Index does not support mutable operations

What is proper way of changing only specific elements of df_days like I have tried to accomplish above?

Edit:

from datetime import timedelta as td

Upvotes: 1

Views: 250

Answers (1)

sgrg
sgrg

Reputation: 1240

Referencing the pandas docs, I found this: http://pandas.pydata.org/pandas-docs/stable/timedeltas.html#id1

You can use this trick along with that code:

tdeltas = pd.TimedeltaIndex([str((0-int(i in hrIdx)))+' days' for i in xrange(len(df_days))])

and then:

df_days = df_days + tdeltas

Note: I'm looking at the docs from the latest version of pandas (0.19).

Upvotes: 2

Related Questions