Python: combine datetime and dataframe

Question

I want to combine a dataframe with datetime. Here is my attempt.

import pandas as pd 
from pandas import DataFrame
import numpy as np

# create df 
df = pd.DataFrame(np.random.uniform(low=10.00, high=32.5, size=(5,2)), columns=['reference', 'target'])

df.head()

reference   target
0   13.559319   24.057471
1   13.101118   19.323373
2   17.295842   18.991576
3   18.123597   13.121553
4   16.328118   28.981584

Then, a timestamp.

import datetime
for a in range(5): # 2 is the number of record
    a = (datetime.date(2010, 1, 15) + datetime.timedelta(a))
    print(a)

2010-01-15
2010-01-16
2010-01-17
2010-01-18
2010-01-19

How to combine both and assign 'date' as the column name for datetime?

you mean final_df=df.assign(date=[(datetime.date(2010, 1, 15) + datetime.timedelta(a)) for a in range(5)]) ?? — anky
– anky, Commented Jul 23, 2019 at 16:08
See the pandas docs for date/time functionality. You might find pd.date_range or pd.to_datetime useful — jonchar
– jonchar, Commented Jul 23, 2019 at 16:13

sjw · Accepted Answer · 2019-07-24 16:30:20Z

4

Here is an option which avoids the need for the loop using pd.date_range:

df['date'] = pd.date_range(start='2010-01-15', end='2010-01-19')

Or:

df['date'] = pd.date_range(start='2010-01-15', freq='1D', periods=len(df))

Either way, result is:

   reference     target       date
0  21.588889  19.881424 2010-01-15
1  10.042584  31.603628 2010-01-16
2  22.734872  11.061602 2010-01-17
3  13.742643  22.620092 2010-01-18
4  19.675036  19.105152 2010-01-19

For the sake of completeness, here is how your original approach would have worked (creating a list of datetimes and then assigning that to a new column of your DataFrame):

dates = [datetime.date(2010, 1, 15) + datetime.timedelta(i) for i in range(len(df))]
df['date'] = dates

This last option is not recommended though as it is likely slower for large datasets, and also you will find it easier to work with the proper Pandas datetime types if you are storing them in a DataFrame, so stick with pd.date_range for this particular use case. Assigning a list of values to a column can be useful in some other non-datetime related cases though.

edited Jul 24, 2019 at 16:30

answered Jul 23, 2019 at 16:12

sjw

6,5512 gold badges30 silver badges41 bronze badges

Sign up to request clarification or add additional context in comments.

7 Comments

anky Over a year ago

nice, also please consider periods=len(df) for the sake of generalizing :)

k.ko3n Over a year ago

Thanks, anky_91, thesilkworm

sjw Over a year ago

@k.ko3n thanks for accepting. By any chance did you accidentally misclick the downvote button while accepting? Seems somebody downvoted and not sure why :)

k.ko3n Over a year ago

@thesilkworm. sorry, it was me. how to fix it?

sjw Over a year ago

@k.ko3n ah I didn't realise that was a rule. Made an edit now which shows how your original approach could have worked (although the other options are still the recommended ones) :)

|

Collectives™ on Stack Overflow

Python: combine datetime and dataframe

1 Answer 1

7 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

7 Comments

Your Answer

Sign up or log in

Post as a guest

Related