CODEWITHSUNDEEP
pythonpandas
Vishakha Lall
Vishakha Lall

Reputation: 1314

Convert hierarchical index to column in Pandas

Here's the input data

df1 = pd.DataFrame( { 
        "author" : ["A","B","A","A","C","B"] , 
        "topic" : ["cat", "dog", "dog", "cat", "dog", "dog"] } )
df1

    author  topic
0   A   cat
1   B   dog
2   A   dog
3   A   cat
4   C   dog
5   B   dog

I'm using group by as follows

g1 = df1.groupby('author')['topic'].value_counts()

author  topic
A       cat      2
        dog      1
B       dog      2
C       dog      1

What I'm looking to achieve is this

author  cat   dog 
A       2     1
B       0     2
C       0     1

Basically, need to convert the second-order of index in hierarchical indexing to columns. How can I do that?

Upvotes: 0

Views: 183

Answers (1)

jezrael
jezrael

Reputation: 862921

Use Series.unstack here:

df = df1.groupby('author')['topic'].value_counts().unstack(fill_value=0)

Another solution with crosstab:

df = pd.crosstab(df1['author'], df1['topic'])

Upvotes: 2

Related Questions