0
        DateTime         car
 2015-04-16 11:57:36     bmw
 2015-04-17 15:32:14     bmw
 2015-04-17 19:13:43     audi
 2015-04-17 05:12:16     porche
 2015-04-17 13:43:31     toyota
 2015-04-15 07:02:20     ferrari

In this dataframe df I need to filter by timestamp -> select one timestamp and for unique car column names I need have to count column which counts unique car names.

Something output looks like this. lets say if we give 2015-04-16 11:57:36

     car        count
     bmw           1
     audi          1
     porche        1
     toyota        1
     ferrari       1

I tried something like this but dont have an idea filter with timestamp. Can anyone help me I got stuck in this part. for car in df['car'].unique(): num = df[df['car'] == car].apply(len)

sai nikhil
  • 1
  • I cannot understand the filter by timestamp: do you need to sum all previously added cars? Or only at the precise timestamp? Is data on the example coherent? Or was just an example? – piertoni Apr 13 '22 at 14:47

1 Answers1

0

You can do a groupby with cumcount like this:

import pandas as pd
d = {'DateTime': ['2015-04-16 11:57:36', '2015-04-17 15:32:14', '2015-04-17 19:13:43','2015-04-17 05:12:16','2015-04-17 13:43:31','2015-04-15 07:02:20'], 'car': ['bmw', 'bmw','audi', 'porche','toyota','ferrari']}
df = pd.DataFrame(data=d)
df["Count"] = df.groupby("DateTime").cumcount() + 1
print(df)

Output

              DateTime      car  Count
0  2015-04-16 11:57:36      bmw      1
1  2015-04-17 15:32:14      bmw      1
2  2015-04-17 19:13:43     audi      1
3  2015-04-17 05:12:16   porche      1
4  2015-04-17 13:43:31   toyota      1
5  2015-04-15 07:02:20  ferrari      1
ACHRAF
  • 341
  • 1
  • 7