[FIXED] Convert column of dates to count of days from start day

Issue

Simple question but I am struggling to find the answer. I am after a simple count of days from the start date. Thanks in advance.

df = pd.DataFrame({'year': [2015, 2015, 2015],
                   'month': [2, 2, 2],
                   'day': [4, 5, 10]})
pd.to_datetime(df)

output

0   2015-02-04
1   2015-02-05
2   2015-02-10

What I want

    Date        Days_from_start
0   2015-02-04  0
1   2015-02-05  1
2   2015-02-10  6

Solution

Create a timedelta by subtracting the reference, then access the days with dt.days:

out = pd.to_datetime(df).to_frame('Date')
out['Days_from_start'] = out['Date'].sub(out['Date'].iloc[0]).dt.days

# or if you want from the oldest day
out['Days_from_start'] = out['Date'].sub(out['Date'].min()).dt.days

Output:

        Date  Days_from_start
0 2015-02-04                0
1 2015-02-05                1
2 2015-02-10                6

Answered By - mozway