Create and fill columns based on conditions

user_14270226 0
Viewed 31

I got a contract DataFrame looking like this :

Customer       CONTRACT_ID      PHASING_DATE    SCHEDULED    ARTICLE_CODE
A              C0218            2021-01-01      21           001
A              COZ19            2021-01-01      23           001
A              IUD80            2021-01-01      43           001
A              PAZO1            2021-02-01      12           002
B              DZAP2            2021-01-01      3            003
B              DZAH8            2021-01-01      4            003
B              FGIG0            2021-03-01      5            003
C              SDFH4            2021-01-01      4            004
C              AZFE3            2021-04-01      54           005
C              DAZJ9            2021-04-01      32           005
..

I would like to create a DataFrame based on the number of active contracts like this :

e.g there's 3 contracts with the same PHASING_DATE, I want 3 columns filled with the scheduled associated to each active contract

CUSTOMER  DATE        ARTICLE_CODE    SCHEDULED_CT_1   SCHEDULED_CT_2      SCHEDULED_CT_3    SCHEDULED_CT_4
A         2021-01-01  001             21               23                  43                0
B         2021-01-01  003             3                4                   0                 0
C         2021-01-01  004             4                0                   0                 0
A         2021-02-01  002             12               0                   0                 0
B         2021-03-01  003             5                0                   0                 0
C         2021-04-01  004             54               32                  0                 0
pandas dataframe group-by python
1 Answers

Use GroupBy.cumcount for counter, add column to DataFrame and pivoting by pivot with replace missing values, rename columns by DataFrame.add_prefix, some data cleaning and last sorting by both columns:

s = df.groupby(['Customer','PHASING_DATE','ARTICLE_CODE']).cumcount()

df = (df.assign(a=s)
        .pivot(['Customer','PHASING_DATE','ARTICLE_CODE'], 'a', 'SCHEDULED')
        .fillna(0)
        .add_prefix('SCHEDULED_CT_')
        .reset_index()
        .rename_axis(None, axis=1)
        .rename(columns={'PHASING_DATE':'DATE'})
        .sort_values(['DATE','Customer']))
print (df)
  Customer        DATE ARTICLE_CODE SCHEDULED_CT_0 SCHEDULED_CT_1  \
0        A  2021-01-01          001             21             23   
2        B  2021-01-01          003              3              4   
4        C  2021-01-01          004              4              0   
1        A  2021-02-01          002             12              0   
3        B  2021-03-01          003              5              0   
5        C  2021-04-01          005             54             32   

  SCHEDULED_CT_2  
0             43  
2              0  
4              0  
1              0  
3              0  
5              0
user_29010020
Related