python - Groupby and any() | all() -

i have following pd.dataframe

in [155]: df1 out[155]:     order_id    acq       date uid 2         3  false 2014-01-03   1 3         4   true 2014-01-04   2 4         5  false 2014-01-05   3 6         7   true 2014-01-08   5 7         8  false 2014-01-08   5 9        10  false 2014-01-10   6 0        11  false 2014-01-11   6 

where each entry order, values order_id, date, uid , acq (indicates whether first order associated uid in dataset).

i trying filter , keep orders placed users have made first order inside time period covered in dataset (i.e. @ least 1 of orders of such users satisfy acq == true).

so, desired output be:

   order_id    acq       date uid 3         4   true 2014-01-04   2 6         7   true 2014-01-08   5 7         8  false 2014-01-08   5 

and have managed reach by:

in [156]: df1.groupby('uid').filter(lambda x: x.acq.any() == true) out[156]:     order_id    acq       date uid 3         4   true 2014-01-04   2 6         7   true 2014-01-08   5 7         8  false 2014-01-08   5 

however, when try find orders placed users have made first order outside time period covered in dataset (i.e. orders should satisfy acq == false) seem lost. have tried this:

in [159]: df1.groupby('uid').filter(lambda x: x.acq.all() == false) out[159]:     order_id    acq       date uid 2         3  false 2014-01-03   1 4         5  false 2014-01-05   3 6         7   true 2014-01-08   5 ## <- order acquisition, therefore orders uid == 5 should filtered out. 7         8  false 2014-01-08   5 9        10  false 2014-01-10   6 0        11  false 2014-01-11   6 

how should go filtering out orders placed users have orders satisfy acq == false?

any ideas appreciated, thanks!

you need first use condition , add all:

print (df1.groupby('uid').filter(lambda x: (x.acq == false).all()))    order_id    acq        date  uid 2         3  false  2014-01-03    1 4         5  false  2014-01-05    3 9        10  false  2014-01-10    6 0        11  false  2014-01-11    6