How to filter df based on corresponding values-CodePudding

I have a df as follows:

data = {'retailer': [2, 2, 2, 2, 2, 5, 5, 5, 5, 5],
        'store': [1, 1, 1, 1, 1, 7, 7, 7, 7, 7],
        'week':[2021110701, 2021101301, 2021100601, 2021092901, 2021092201, 2021110701, 2021101301, 2021100601, 2021092901, 2021092201],
        'isPeriod': [False, True, False, False, False, False, False, True, False, False],
        'quadId': [2021112804, 2021103104, 2021103104, 2021103104, 2021103104, 2021100304, 2021100304, 2021103104, 2021103104, 2021103104]
       }
data = pd.DataFrame(data)

I would like to located where 'isPeriod' == True, get the corresponding 'quadId' values for where 'isPeriod' is True and then filter my entire dataframe to only have that corresponding 'quadId'.

For example, in my df we can see that in the second row, 'isPeriod' is True and the corresponding 'quadId' is 2021103104. So I would like my filtered df to only contain the rows where 'quadId' is 2021103104.

In this case my example filtered df would look like:

data = {'retailer': [2, 2, 2, 2, 5, 5, 5],
        'store': [1, 1, 1, 1, 7, 7, 7],
        'week':[2021101301, 2021100601, 2021092901, 2021092201, 2021100601, 2021092901, 2021092201],
        'isPeriod': [True, False, False, False, True, False, False],
        'quadId': [2021103104, 2021103104, 2021103104, 2021103104, 2021103104, 2021103104, 2021103104]
       }
data = pd.DataFrame(data)

Is there a way I can do this? Thanks! (Also wherever if there are multiple True values for isPeriod, the quadId's will all be the same for them)

CodePudding user response：

Use isin to check for existence, then loc:

valid_quarters = data.loc[data.isPeriod, 'quadId']
data[data['quadId'].isin(valid_quarters)]

Output:

   retailer  store        week  isPeriod      quadId
1         2      1  2021101301      True  2021103104
2         2      1  2021100601     False  2021103104
3         2      1  2021092901     False  2021103104
4         2      1  2021092201     False  2021103104
7         5      7  2021100601      True  2021103104
8         5      7  2021092901     False  2021103104
9         5      7  2021092201     False  2021103104