CODEWITHSUNDEEP
pythondatetimepandas
ccsv
ccsv

Reputation: 8659

python pandas Slicing datetime dates by number of rows

Lets say I have a time series

import pandas as pd
from numpy.random import randn

dates = pd.date_range('12/31/2014', periods=10)
df = pd.DataFrame(randn(10, 4), index=dates, columns=['A', 'B', 'C', 'D'])

Given a date such as d ='1/5/2015' how would I select the rows two days after d (days = 1/6/2015, 1/7/2015) and two days before d (days = 1/4/2015, 1/3/2015)? Is there a way to do this to ignore missing data from either weekends or holidays?

Upvotes: 2

Views: 781

Answers (2)

elyase
elyase

Reputation: 40973

You can do it like this:

from pandas.tseries.offsets import BDay

d = pd.Timestamp('1/5/2015')
two_bdays_before = d - BDay(2)   # business days
two_bdays_later = d + BDay(2)

Then to access all days between two_bdays_before and two_bdays_later:

>>> df[two_bdays_before:two_bdays_later]]
                   A         B         C         D
2015-01-01  0.741045 -0.051576  0.228247 -0.429165
2015-01-02 -0.312247 -0.391012 -0.256515 -0.849694
2015-01-03 -0.581522 -1.472528  0.431249  0.673033
2015-01-04 -1.408855  0.564948  1.019376  2.986657
2015-01-05 -0.566606 -0.316533  1.201412 -1.390179
2015-01-06 -0.052672  0.293277 -0.566395 -1.591686
2015-01-07 -1.669806  1.699540  0.082697 -1.229178

Upvotes: 2

unutbu
unutbu

Reputation: 879361

df.index.get_loc(d) returns an integer index corresponding to the date represented by the date string d.

You can then use that integer index to select 2 rows before or after d in df:

import pandas as pd
import numpy as np

dates = pd.date_range('12/31/2014', periods=10)
df = pd.DataFrame(np.random.randn(10, 4), index=dates, columns=['A', 'B', 'C', 'D'])
d = '1/5/2015'

idx = df.index.get_loc(d)
print(df.iloc[idx+1:idx+3])
#                    A         B         C         D
# 2015-01-06  1.211569  1.766432  0.153963  1.101142
# 2015-01-07  0.018377  0.112825  0.347711 -1.400145

print(df.iloc[idx-2:idx])
#                    A         B         C         D
# 2015-01-03 -0.507956 -1.389623 -0.092228 -0.104655
# 2015-01-04  0.206824  1.226987  0.253424 -0.529778

Upvotes: 1

Related Questions