切片时间序列

Question

Let's say I have a time series "TS" with the range of months. How can I get a sub-time series from TS with for example the range of one day.

I tried this code :

subts = pd.Series(TS,index=pd.to_datetime(['2015-09-26','2015-09-27']))

But I get this error :

ValueError: Wrong number of items passed 472, placement implies 2

What I understood is, the method I chose matches every value from TS (472rows) an the time range I gave in the constructor (ie : ['2015-09-26','2015-09-27'])

Is there a way to really slice a time series? Just extract a part of it within a given time range?

Answer 1

I think you can use selecting by [] , see also indexing :

subts = TS['2015-09-26':'2015-09-27']

Or:

subts = TS.loc['2015-09-26':'2015-09-27']

Sample:

np.random.seed(123)
TS = pd.Series(np.random.randint(10, size=10), pd.date_range('2015-09-24', periods=10))
print (TS)
2015-09-24    2
2015-09-25    2
2015-09-26    6
2015-09-27    1
2015-09-28    3
2015-09-29    9
2015-09-30    6
2015-10-01    1
2015-10-02    0
2015-10-03    1
Freq: D, dtype: int32

subts = TS['2015-09-26':'2015-09-27']
print (subts)
2015-09-26    6
2015-09-27    1
Freq: D, dtype: int32