CODEWITHSUNDEEP
pythonpandas
gontrollez
gontrollez

Reputation: 6548

Filter rows by condition applied to indices

I have a DataFrame with dates as indices:

            VL
2018-02-05  101.56093
2018-12-31  95.87728
2019-01-04  96.29820
2019-01-11  97.23475
2019-01-18  98.39828
2019-01-25  98.66896
2019-01-31  99.12407
2019-02-01  99.13224
2019-02-08  99.06382
2019-02-15  99.79966

I need to filter the rows so that, for each row with date D, keep it if the row with D-7 exists in the DataFrame.

Example:

I've implemented this already using a loop but I'm wondering if there is a more pandas oriented way of doing this kind of filtering.

Upvotes: 0

Views: 32

Answers (1)

Scott Boston
Scott Boston

Reputation: 153500

IIUC, you can use pd.Timedelta and isin:

df[(df['date'] - pd.Timedelta(days=7)).isin(df['date'])]

Output:

        date        VL
3 2019-01-11  97.23475
4 2019-01-18  98.39828
5 2019-01-25  98.66896
7 2019-02-01  99.13224
8 2019-02-08  99.06382
9 2019-02-15  99.79966

If date is in the index use this:

df[(df.index - pd.Timedelta(days=7)).isin(df.index)]

Output:

                  VL
date                
2019-01-11  97.23475
2019-01-18  98.39828
2019-01-25  98.66896
2019-02-01  99.13224
2019-02-08  99.06382
2019-02-15  99.79966

Upvotes: 2

Related Questions