Python Forum

I want to sort column 'PERIOD_START_TIME' from new date to old date.

df_pivot = pd.pivot_table (df, values = 'NTM NR RET NSA SN Total Drops',index = ['CELLCU'], columns = 'PERIOD_START_TIME', aggfunc = 'sum')

Output:
CELLCU	10/13/2022	10/14/2022	10/15/2022	10/16/2022	10/17/2022
A1WA0944C11	580	384	590	644	1464
A1WA0944C21	2065	3649	2687	2220	3205
A1WA0944C31	1874	1829	1901	2617	3111
A1WA6372A11	1327	1507	404	541	1478
A1WA6372A21	2559	35351	16881	9500	2203
A1WA6372A31	748	7422	563	914	1086
A1WA6374A11	84	93	1828	1339	414
A1WA6374A21	119	463	271	2245	628
A1WA6374A31	1532	6173	36461	78170	15273
A1WA6375A11	98	93	223	188	120

If this is a question about pivot tables, please provide a sample of the original data and describe the result you want after pivoting. Your problem might be solved by sorting the original dataframe before pivoting.

If you want to sort your pivoted dataframe columns:

from io import StringIO
import pandas as pd
from datetime import datetime

# Reproduce pivoted dataframe
csv_data= StringIO("""
CELLCU 10/13/2022 10/14/2022 10/15/2022 10/16/2022 10/17/2022
A1WA0944C11 580 384 590 644 1464
A1WA0944C21 2065 3649 2687 2220 3205
A1WA0944C31 1874 1829 1901 2617 3111
A1WA6372A11 1327 1507 404 541 1478
A1WA6372A21 2559 35351 16881 9500 2203
A1WA6372A31 748 7422 563 914 1086
A1WA6374A11 84 93 1828 1339 414
A1WA6374A21 119 463 271 2245 628
A1WA6374A31 1532 6173 36461 78170 15273
A1WA6375A11 98 93 223 188 120""")

df = pd.read_csv(csv_data, sep=" ")

# Sort columns[1:].  Month/Day/Year is not good for sorting.  Use datetime key.
columns = list(df.columns)
dates = sorted(columns[1:], key=lambda x:datetime.strptime(x, "%m/%d/%Y"), reverse=True)
df = df.reindex([columns[0]] + dates, axis=1)
print(df)