CODEWITHSUNDEEP
pythondatetimepandas
Amelio Vazquez-Reina
Amelio Vazquez-Reina

Reputation: 96350

datetime range indexing: datetimes that may not be in the index?

I have a dataframe indexed by datetime objects: 

In <10>: all_data.head().index
Out<10>: 
Index([2014-04-23, 2014-04-13, 2014-04-15, 2014-04-30, 2014-04-06], dtype='object')

and two timestamps:

In <11>: d1
Out<11>: datetime.datetime(2014, 3, 24, 0, 0)

In <12>: d2
Out<12>: datetime.datetime(2014, 4, 6, 0, 0)

I would like to index a column base don the d1:d2 range. Note that d1 or d2 may not be in the index. How can I do this in Pandas?

I tried:

all_data.loc[d1:d2,:]

but I get: start bound[2014-03-24 00:00:00] is not the [index]

Upvotes: 3

Views: 1634

Answers (1)

Karl D.
Karl D.

Reputation: 13757

Well, if you make the index a DateTimeIndex, partial string indexing should work:

print df
print df.index

            x1  x2
date              
2014-04-23   1   2
2014-04-13   2   4
2014-04-15   3   6
2014-04-30   4   8
2014-04-06   5  10

[5 rows x 2 columns]

<class 'pandas.tseries.index.DatetimeIndex'>
[2014-04-23, ..., 2014-04-06]

Then you can use partial string slicing:

print df['2014-03-24':'2014-04-06']

            x1  x2
2014-04-06   5  10

or

print df.ix['2014-03-24':'2014-04-13',:]

            x1  x2
date              
2014-04-13   2   4
2014-04-06   5  10

Upvotes: 3

Related Questions