How to create a dataframe from another dataframe based on GroupBy condition

Question

I don't know how can i create a dataframe based on another dataframe using a groupby conditions. For example, i have a dataframe that if i apply the function:

flights_df.groupby(by='DepHour')['Cancelled'].value_counts()

I obtain something like this

DepHour  Cancelled
0.0      0             20361
         1                 7
1.0      0              5857
         1                 4
2.0      0              1850
         1                 1
**3.0      0               833**
4.0      0              3389
         1                 1
5.0      0            148143
         1                24

As can be seen, for DepHour == 3.0 there's no cancelled flights.

Using the same dataframe that i used to generate this output i want to create another dataframe containing only of values where there's no cancelled flighs for DepHour. In this case, the output will be a dataframe containing only values of DepHour == 3.0 .

I know that i can use mask, but it allows only filter values where cancelled == 0 (ie all other values for where DepHour cancelled == 0 are included).

Thanks and sorry for my bad english!

Answer 1

There might be a cleaner way (probably without using groupby twice) but this should should work:

flights_df.groupby('DepHour') \
.filter(lambda x: (x['Cancelled'].unique()==[0]).all()) \
.groupby('DepHour')['Cancelled'].value_counts()