Exclude groups containing just one row instance

Question

I have a pandas table like below with groupby applied to get Groups 0, 1 and 2 as follows:

Group 0	Group 1	Group 2	Count
A	X1	577.5000	6
		894.8700	2
	X2	2697.3100	48
		2697.3100	1
	B3	2697.3100	30
B	RO2	92.46	3
C	C12	34.2700	9
		39.2700	3
D	X55	4.1251	9

How do I filter out rows having just one group variation i.e. Below is how the final DF should look like:

Group 0	Group 1	Group 2	Count
A	X1	577.5000	6
		894.8700	2
	X2	2697.3100	48
		2697.3100	1
	B3	2697.3100	30
C	C12	34.2700	9
		39.2700	3

Asked By: Krishna

||

Source

Answer 1

You can keep groups where at least one value from Group 0 is duplicated:

out = df.loc[df.index.get_level_values('Group 0').duplicated(keep=False)]
print(out)

# Output
                         Count
Group 0 Group 1 Group 2       
A       X1      577.50       6
                894.87       2
        X2      2697.31     48
                2697.31      1
        B3      2697.31     30
C       C12     34.27        9
                39.27        3

Answered By: Corralien

Answer 2

Use groupby.size and boolean indexing:

Only considering the first group:

out = df[df.groupby(['Group 0'])['Group 0'].size().gt(1)]

Considering all groups:

out = df[df.groupby(['Group 0', 'Group 1', 'Group 2'])['Group 0'].size().gt(1)]

Answered By: mozway

Exclude groups containing just one row instance

Question:

Answers: