Sum different rows in a data frame based on multiple conditions-CodePudding

I have the following data frame:

dataset = {
     'String': ["ABCDEF","HIJABC","ABCHIJ","DEFABC"],
    'Bool':[True,True,False,False],
    'Number':[10,20,40,50]}

df = pd.DataFrame(dataset)



      String   Bool  Number
0  ABCDEF   True      10
1  HIJABC   True      20
2  ABCHIJ  False      40
3  DEFABC  False      50

I would like to sum the rows of the column Number where Bool is False to the rows where Bool is True:

The rows can be matched and summed together if the reverse of String of one row is not equal to the String of the row.

In this case ABCHIJ where bool is False is not equal to the reverse of ABCDEF so I sum the numbers: 10 40.
HIJABC where Bool is True is summed to DEFABC where Bool is False the outcome is 70

Expected output:

    String   Bool  Number
0  ABCDEF   True      50
1  HIJABC   True      70
2  ABCHIJ  False      40
3  DEFABC  False      50

I hope my explanation was good enough, is there a way to achieve the above outcome ?

CodePudding user response：

One way is like this:

df_true = df[df['Bool'] == True]
df_false = df[df['Bool'] == False]

for i in df_false['String']:
   idx = df_true[df_true['String'] != (i[3:]   i[:3]) ].index[0]
   current_num = df.loc[df.index == idx, 'Number'].values[0]
   added_num = df[df['String'] == i]['Number'].values[0]
   df.loc[df.index == idx, 'Number'] =  current_num   added_num

I hope it helps