How to add counter columns to get the frequency of items in dataframe-CodePudding

i have such dataframe:

product1	product2	product3	product4	product5	product6
straws	orange	melon	chair	bread	milk
melon	milk	book	coffee	cake	tea
bread	bananas	juice	chair	book	straws

I wish to add 6 counting items columns along the index and counts per rows the number of times they appear in the dataframe.

CountProduct1	CountProduct2	CountProduct3	Countproduct4	Countproduct5
1	1	1	1	1
2	2	1	1	1
2	1	1	2	1

Thanks for your help.

CodePudding user response：

Given the modified explanation, here is a way to do it.

1. cumulative counts

cnt = df.apply(lambda g: g.value_counts(), axis=1).fillna(0).astype(int).cumsum()
>>> cnt
   bananas  book  bread  cake  chair  coffee  juice  melon  milk  orange  straws  tea
0  0        0     1      0     1      0       0      1      1     1       1       0  
1  0        1     1      1     1      1       0      2      2     1       1       1  
2  1        2     2      1     2      1       1      2      2     1       2       1

2. replace each item by the corresponding count, row by row

out = pd.DataFrame([
    r.map(c) for (_, r), (_, c) in zip(df.T.items(), cnt.T.items())
]).set_axis([f'Count{k}' for k in df.columns], axis=1)

>>> out
   Countproduct1  Countproduct2  Countproduct3  Countproduct4  Countproduct5  Countproduct6
0  1              1              1              1              1              1            
1  2              2              1              1              1              1            
2  2              1              1              2              2              2