suppose have data frame 3 columns as
date amount type 0 20160101 50 apple 1 20160101 50 apple 2 20160101 50 banana 3 20160102 30 apple 4 20160102 50 apple 5 20160102 40 banana 6 20160102 40 banana what want count frequency of occurrence of row using columns, , result should like
date amount type times 20160101 50 apple 2 20160101 50 banana 1 20160102 30 apple 1 20160102 50 apple 1 20160102 40 banana 2 my code like
df out[23]: date amount type 0 20160101 50 apple 1 20160101 50 apple 2 20160101 50 banana 3 20160102 30 apple 4 20160102 50 apple 5 20160102 40 banana 6 20160102 40 banana p=df.pivot_table('amount','date','type') p out[27]: type apple banana date 20160101 50 50 20160102 40 40
here's hack answer. feel there needs more direct way though
df['times'] = 1 df.groupby(['date', 'amount', 'type'], as_index=false).sum() edit:
found second solution (you need rename column
df.groupby(['date','type']).amount.value_counts().reset_index() 