I have a pandas dataframe with the following content (X represents some value)
| personal id | time point | label |
|---|---|---|
| 1 | 1 | X |
| . | . | . |
| 1 | 12 | X |
| 2 | 400 | X |
| . | . | . |
| 2 | 411 | X |
Basically, for every person, I have labels for 12 consecutive time entries (not always from 1 to 12). What I want is the following form
| personal id | label (first time point) | label (seconde time point) | ... | label (twelfth time point) |
|---|---|---|---|---|
| 1 | X | X | ... | X |
| 2 | X | X | ... | X |
How can I achieve this? Since I want to ignore the time point numbers and only want the twelve labels next to each other, pivot tables doesn't seem to do the job.
CodePudding user response：
Use GroupBy.cumcount for counter by personal id and then use DataFrame.pivot:
df = df.assign(g = df.groupby('personal id').cumcount()).pivot('personal id','g','label')