Python Pandas：如何排除任何值重复超过 n 次的行

Question

I'm pretty green when it comes to Python so my apologies if this is an obvious question.

I have a dataframe that has 8 columns. For each row the first four columns are single names. The following four columns are locations that each one of those names are associated with. Here's an example.

name1	name2	name3	name4	loc1	loc2	loc3	loc4
Joe	Dave	Aaron	Alex	NYC	CHI	ANN	FAL
Erica	Alana	Steve	Blake	JAX	MIA	JAX	JAX
Stacy	Tom	Nancy	Steph	SAC	SFR	DAL	DAL

All I want to do is take that dataframe and create a new one that shows all the same information but excludes any rows that have more than two of the same location in the last 4 columns. and then the result needs to have the index reset. So the result of the example above would be:

name1	name2	name3	name4	loc1	loc2	loc3	loc4
Joe	Dave	Aaron	Alex	NYC	CHI	ANN	FAL
Stacy	Tom	Nancy	Steph	SAC	SFR	DAL	DAL

I was trying to make it work with a combination of apply, groupby and count but could not get it to work right. I feel like there's a simple solution.

Many thanks!

Answer 1

You can use nunique on each row:

df[df[['loc1', 'loc2', 'loc3', 'loc4']].apply(lambda x: len(x) - x.nunique() < 2, axis=1)].copy().reset_index()

Answer 2

Try this:

df[df.filter(like = 'loc').nunique(axis = 1) > 2]

Output:

   name1 name2  name3  name4 loc1 loc2 loc3 loc4
0    Joe  Dave  Aaron   Alex  NYC  CHI  ANN  FAL
2  Stacy   Tom  Nancy  Steph  SAC  SFR  DAL  DAL

Using filter with like = 'loc' to limit the dataframe columns to just the last four or the loc columns, then get the number of unique values with axis=1 using nunique , the create a boolean series and with boolean filtering we can select correct rows.