CODEWITHSUNDEEP
pythonpandasnumpy
yaw
yaw

Reputation: 309

Splitting dataframe into groups on condition

I've got dataframe with below structure:

Id.    Value.

 1. 10
 2. 12
 3. 14
 4. 16
 5. 30
 6. 32
 7. 35
 8. 36
 9. 60
 10. 61
 11. 63
 12. 67
 13. 70

And I want to split it into number of groups using a conditional statement, which is comparing subsequent rows in my dataframe, What is the best way to do it?

Upvotes: 0

Views: 244

Answers (1)

Scott Boston
Scott Boston

Reputation: 153500

IIUC, you try could this:

s = pd.Series([10,12,14,16,30,32,35,36,60,61,63,67,70])
s.groupby((s.diff() >= 10).cumsum()).apply(list)

Output:

0        [10, 12, 14, 16]
1        [30, 32, 35, 36]
2    [60, 61, 63, 67, 70]
dtype: object

For a dataframe:

df = pd.DataFrame({'id':np.arange(1,14),
                  'Value':[10,12,14,16,30,32,35,36,60,61,63,67,70]})

df.groupby(df['Value'].diff().ge(10).cumsum())['Value'].apply(list).tolist()

Output:

[[10, 12, 14, 16], [30, 32, 35, 36], [60, 61, 63, 67, 70]]

Upvotes: 1

Related Questions