I have a table (pivot table) like this:
value B N S date
date
2020-12-31 1 11 0 2020-12-31
2021-01-01 3 80 0 2021-01-01
2021-01-02 4 99 0 2021-01-02
2021-01-03 3 78 0 2021-01-03
2021-01-04 0 50 0 2021-01-04
and now I want to make it like this by adding a new column:
value B N S date yesterday_B
date
2020-12-31 1 11 0 2020-12-31 NaN
2021-01-01 3 80 0 2021-01-01 1
2021-01-02 4 99 0 2021-01-02 3
2021-01-03 3 78 0 2021-01-03 4
2021-01-04 0 50 0 2021-01-04 3
is there and pandas function for it? what should I do?? (after all, I want to use it as a feature (consider B and the value for its previous day) in the machine learning model. can I do it without adding a new column?
CodePudding user response:
Use Series.shift()
df['yesterday_B'] = df['B'].shift()