Date TIme in Python for data scientists
python date datatype- datetime
pandas date datatype- Timestamp
1) creating pandas datframe with string dates value
## datetime
date_data= pd.DataFrame(np.random.randint(10,100, (4,3)), columns=['A','B','C'])
dates= ['2 june 2013', '5 Aug 2015', '2015-07-09', '7/12/2014']
date_data.index= dates
4) plotting time series
pandas date datatype- Timestamp
1) creating pandas datframe with string dates value
## datetime
date_data= pd.DataFrame(np.random.randint(10,100, (4,3)), columns=['A','B','C'])
dates= ['2 june 2013', '5 Aug 2015', '2015-07-09', '7/12/2014']
date_data.index= dates
A | B | C | |
---|---|---|---|
2 june 2013 | 52 | 61 | 89 |
5 Aug 2015 | 21 | 69 | 89 |
2015-07-09 | 88 | 23 | 13 |
7/12/2014 | 43 | 39 | 21 |
creating string into pandas datetime format-
date_data.index= pd.to_datetime(date_data.index)
date_data
A | B | C | |
---|---|---|---|
2013-06-02 | 59 | 93 | 77 |
2015-08-05 | 33 | 15 | 28 |
2015-07-09 | 63 | 19 | 25 |
2014-07-12 | 29 | 36 | 92 |
2) Time difference in pandas-
pd.Timestamp('12.03.2019')- pd.Timestamp('12 Aug 2018')
Timedelta('478 days 00:00:00')
3) Return a fixed frequency DatetimeIndex -
cum_array= pd.DataFrame({'colA': np.random.randint(1,5,9).cumsum(), 'colB': np.random.randint(-4,10,9)})
dates=pd.date_range('10-10-2019','15-10-2019', periods=9)
cum_array.index= dates
cum_array.diff() # provides difference
import matplotlib.pyplot as plt
%matplotlib inline
cum_array.plot()
In [ ]:
Comments
Post a Comment