2

this is a basic quesiton and easy to do in excel but have not an idea in python and every example online uses groupby with multiple names in the name column. So, all I need is a row value of weights from a single column. Suppose I have data that looks like this:

        name    value
0       A       45
1       B       76
2       C       320
3       D       210

The answer should look like this:

0      name     value  weights
1      A        45     0.069124
2      B        76     0.116743
3      C        320    0.491551
4      D        210    0.322581

thank you,

user2100039
  • 1,280
  • 2
  • 16
  • 31

2 Answers2

2

Use GroupBy.transform for repeat sum per groups, so possible divide original column:

print (df.groupby('name')['value'].transform('sum'))
0    651
1    651
2    651
3    651
Name: value, dtype: int64

df['weights'] = df['value'].div(df.groupby('name')['value'].transform('sum'))
print (df)
  name  value   weights
0    A     45  0.069124
1    A     76  0.116743
2    A    320  0.491551
3    A    210  0.322581

EDIT:

df['weights'] = df['value'].div(df['value'].sum())
print (df)
  name  value   weights
0    A     45  0.069124
1    B     76  0.116743
2    C    320  0.491551
3    D    210  0.322581
jezrael
  • 822,522
  • 95
  • 1,334
  • 1,252
  • 1
    sorry, my data in the 'name' column has different names for each row - example is not correct. can you take a look after the edits? – user2100039 Dec 15 '21 at 06:29
  • @user2100039 - So need divide sum of all column by column like in `EDIT` ? – jezrael Dec 15 '21 at 06:32
2

You can also groupby 'name' and then apply a function that divides each value by its group sum:

df['weights'] = df.groupby('name')['value'].apply(lambda x: x / x.sum())

Output:

  name  value   weights
0    A     45  0.069124
1    A     76  0.116743
2    A    320  0.491551
3    A    210  0.322581

For new data:

df['weights'] = df['value'] / df['value'].sum()

  name  value   weights
0    A     45  0.069124
1    B     76  0.116743
2    C    320  0.491551
3    D    210  0.322581