Convert the following table:
| Type | s | h |
|---|---|---|
| A | 1 | 4 |
| A | 2 | 5 |
| B | 4 | 7 |
| B | 5 | 2 |
into this:
| Type | s_1 | h_1 | s_2 | h_2 |
|---|---|---|---|---|
| A | 1 | 4 | 2 | 5 |
| B | 4 | 7 | 5 | 2 |
CodePudding user response：
In your case doing the cumcount with groupby get the key then we can pivot
out = df.assign(key = df.groupby('Type').cumcount() 1).pivot(index='Type',columns='key').sort_index(level=1,axis=1)
out.columns = out.columns.map('{0[0]}_{0[1]}'.format)
out = out.reset_index()
out
Out[581]:
Type h_1 s_1 h_2 s_2
0 A 4 1 5 2
1 B 7 4 2 5
CodePudding user response：
df = df.groupby('Type').agg(list)
df[['s_1', 's_2']] = df.apply(lambda x: x['s'], axis=1, result_type='expand')
df[['h_1', 'h_2']] = df.apply(lambda x: x['h'], axis=1, result_type='expand')
df = df.drop(['s', 'h'], axis=1).reset_index()
print(df)
Output:
Type s_1 s_2 h_1 h_2
0 A 1 2 4 5
1 B 4 5 7 2