How can a I drop duplicate rows for a dataframe based on the filter or condition of another column?-CodePudding

I have a large dataframe which is in long format and can be created below:

import pandas as pd
df = pd.DataFrame({'period':['2021-01-01','2021-02-01','2021-03-01','2021-03-01','2021-04-01','2021-04-01'],
                    'indi':['pop','vacced','tot_num_cases','tot_num_cases','pop','pop'],
                    'value':[10000,200,8999,8999,27000,27000]})

I want to drop duplicate rows based on the condition below:

df[df['indi'] == 'tot_num_cases'].drop_duplicates(keep="last")

but only on the rows which match the condition. How do that without dropping all duplicate rows of the dataframe. Result would look like:

CodePudding user response：

Split the condition to two conditions, duplicate value in indi column and the value is 'tot_num_cases'

df = df[~(df.duplicated(subset='indi', keep='last') & df['indi'].eq('tot_num_cases'))]
print(df)

Output

       period           indi  value
0  2021-01-01            pop  10000
1  2021-02-01         vacced    200
3  2021-03-01  tot_num_cases   8999
4  2021-04-01            pop  27000
5  2021-04-01            pop  27000

CodePudding user response：

You can select tot_num_cases, drop duplicates and concat with the other rows:

import pandas as pd
df = pd.DataFrame({'period':['2021-01-01','2021-02-01','2021-03-01','2021-03-01','2021-04-01','2021-04-01'],
                    'indi':['pop','vacced','tot_num_cases','tot_num_cases','pop','pop'],
                    'value':[10000,200,8999,8999,27000,27000]})
#        period           indi  value
# 0  2021-01-01            pop  10000
# 1  2021-02-01         vacced    200
# 2  2021-03-01  tot_num_cases   8999
# 3  2021-03-01  tot_num_cases   8999
# 4  2021-04-01            pop  27000
# 5  2021-04-01            pop  27000


df = pd.concat([df.loc[df['indi']=='tot_num_cases'].drop_duplicates(keep="last"), 
                df.loc[df['indi']!='tot_num_cases']])

#        period           indi  value
# 3  2021-03-01  tot_num_cases   8999
# 0  2021-01-01            pop  10000
# 1  2021-02-01         vacced    200
# 4  2021-04-01            pop  27000
# 5  2021-04-01            pop  27000