I have the following:
count = a[(a.type == 'exact') & (a.score == 1.0)].groupby('id').count()['type']
countbyid = pd.DataFrame(data = count)
Which gives me the following:
id type
102116 578
1256 1211
126805 215
31560 12
329401 2310
49923 375
9691 2409
When I try to add column names to countbyid = pd.DataFrame(data = count, columns = ['customer', 'department']) NO data is returned, only the column names:
customer department
With your given approach, you are inserting the dataframe which already has column names and when you assign the columns inside, no values against those columns are assigned, hence you receive a null dataframe.
So two approaches to tackle this problem -
First:
countbyid = pd.DataFrame(data = countbyid.values)
countbyid.columns = ['customer', 'department']
Second (as mentioned in comment by ScottBoston):
countbyid = pd.DataFrame(data= countbyid.values, columns = ['customer', 'department'])
Based on your code, I assume that a is a dataframe, you just need to reset index and rename columns. No need to call DataFrame constructor again.
a[(a.type == 'exact') & (a.score == 1.0)].groupby('id').type.count().reset_index()\
.rename({'id':'customer', 'type':'department'})
