CODEWITHSUNDEEP
pythonpandasdictionary
Juan C
Juan C

Reputation: 6132

How to map dict with list of indexes to new variable

So, today I'm struggling with the following situation: I have a dictionary where each key is a cluster number, and each value is a list comprised of the index numbers associated to that cluster:

dic={'0':[0,1,2],'1':[3,4,5]}

And my DataFrame looks like this:

index    col
 0       foo
 1       foo
 2       foo
 3       foo
 4       foo
 5       foo

I thought the following would work:

df['cluster']=df.index.map(dic)

But it maps to the keys and not the values, bringing the following:

index    col   cluster
 0       foo   [0,1,2]
 1       foo   [3,4,5]
 2       foo     nan
 3       foo     nan
 4       foo     nan
 5       foo     nan

And what I want is:

index    col   cluster
 0       foo     0
 1       foo     0
 2       foo     0
 3       foo     1
 4       foo     1
 5       foo     1

Is there any other way than reversing my dictionary to map this?

Upvotes: 1

Views: 248

Answers (1)

BENY
BENY

Reputation: 323306

Check with flatten your dict 

df.index.map({y : x[0] for x in dic.items() for y in x[1]})
Out[379]: Index(['0', '0', '0', '1', '1', '1'], dtype='object')
#df['cluster']=df.index.map({y : x[0] for x in dic.items() for y in x[1]})

Upvotes: 2

Related Questions