CODEWITHSUNDEEP
pythonpandas
user9270170
user9270170

Reputation:

How to split the dataframe into sub-dataframes according to different groups？

The data in my csv likes this:

staff_id    clock_time  device_id   latitude    longitude
1001    2020/9/20 7:26  d_1 24.48237852 118.1558955
1001    2020/9/20 5:30  d_1 24.59689407 118.0863806
1001    2020/9/18 4:17  d_2 24.59222786 118.0955275
1001    2020/9/16 3:33  d_2 24.59208312 118.0957197
1001    2020/9/15 8:34  d_2 24.59732974 118.0859644
1001    2020/9/14 4:43  d_2 25.68714724 119.3918519
1002    2020/9/13 1:17  d_3 24.58618235 118.1065715
1002    2020/9/11 9:20  d_3 24.63024186 118.0667598
1002    2020/9/10 7:22  d_3 24.48287497 118.1542957
1002    2020/9/7 5:38   d_4 25.07601853 118.7335211
1003    2020/9/5 5:44   d_6 24.59803941 118.0863071
1003    2020/9/4 7:37   d_6 24.48285023 118.1545752
1003    2020/9/3 2:38   d_6 24.6381382  118.0677933
1003    2020/8/31 6:43  d_7 24.49278011 118.1395677
1003    2020/8/30 11:41 d_7 24.59205252 118.0955596
1003    2020/8/29 3:35  d_7 24.51817637 118.1764342
1003    2020/8/28 5:05  d_7 24.59603175 118.0846872
1003    2020/8/27 2:55  d_8 26.39899424 117.7866387
1003    2020/8/26 7:45  d_8 26.39900029 117.7866379
1003    2020/8/26 3:09  d_8 26.40672436 117.8008659
1003    2020/8/26 0:26  d_8 26.89169118 117.1612365
1003    2020/8/25 9:38  d_8 26.89764297 117.1760012
1003    2020/5/19 8:29  d_8 24.47420087 118.1085551
1003    2020/5/18 9:06  d_8 24.473124   118.1705641
1003    2020/5/16 7:54  d_8 24.5101858  117.8954614

I want to split the dataframe into sub-dataframes according to staff_id and device_id, and put these sub-dataframe into a list,for example:

sub-dataframe1 is:

1001    2020/9/20 7:26  d_1 24.48237852 118.1558955
1001    2020/9/20 5:30  d_1 24.59689407 118.0863806

sub-dataframe2 is:

1001    2020/9/18 4:17  d_2 24.59222786 118.0955275
1001    2020/9/16 3:33  d_2 24.59208312 118.0957197
1001    2020/9/15 8:34  d_2 24.59732974 118.0859644
1001    2020/9/14 4:43  d_2 25.68714724 119.3918519

sub-dataframe3 is:

1002    2020/9/13 1:17  d_3 24.58618235 118.1065715
1002    2020/9/11 9:20  d_3 24.63024186 118.0667598
1002    2020/9/10 7:22  d_3 24.48287497 118.1542957

So on.

How to do this?

My code:

import pandas as pd
df = pd.read_csv(r'for_test.csv', sep=',', encoding='utf-8')
gb = df.groupby(['staff_id','device_id'])

Upvotes: 0

Views: 114

Answers (2)

Yashar
Yashar

Reputation: 832

If you want to extract separate DataFrames, you just need to iterate over your own solution:

import pandas as pd
df = pd.read_csv(r'for_test.csv', sep=',', encoding='utf-8')
gb = df.groupby(['staff_id','device_id'])
l = []
for i in gb.indices:
    df = pd.DataFrame(gb.get_group(i))
    l.append(df)

At the end, you will have separate data frames based on each staff_id/device_id pair in the list "l". Just have to add I think there are cleaner ways to iterate over the GroupBy object, but this will do for now.

Upvotes: 1

Vivs
Vivs

Reputation: 475

Try this.

import pandas as pd
df = pd.read_csv(r'for_test.csv', sep=',', encoding='utf-8')
gb = df.groupby('device_id')
#print(gb.first())
#print the groups for example 'd_1'
print(gb.get_group('d_1') )
#to convert to list
print(gb.get_group('d_1').values.tolist())

Upvotes: 0

Related Questions