CODEWITHSUNDEEP
pythonpandas
Watt
Watt

Reputation: 3164

Transform non numeric rows into columns with custom column name

I have a dataframe which looks something like this:

    df = pd.DataFrame([['7690d79f','Pos'],['7690d79f','Neg'],['7690d79f','Pos'],['7690d79f','Neu'],\
                       ['6690d79f','Pos'],['6690d79f','Neg'],['6690d79f','Pos'],['6690d79f','Neu']]\
                       ,columns=['id', 'vote'])

which is:

    id         vote
0   7690d79f    Pos
1   7690d79f    Neg
2   7690d79f    Pos
3   7690d79f    Neu
4   6690d79f    Pos
5   6690d79f    Neg
6   6690d79f    Pos
7   6690d79f    Neu

I want to pivot rows to columns such that, I get a resultDF as

id  vote_1 vote_2 vote_3 vote_4
7690d79f Pos  Neg Pos Neu
6690d79f Pos  Neg Pos Neu

If it helps, there is no more than 4 votes per ids.

Though this is similar to

  1. Python Pandas: Convert Rows as Column headers

  2. Convert row to column in Python Pandas

But unfortunately, none of them works for me.

Upvotes: 1

Views: 227

Answers (1)

Scott Boston
Scott Boston

Reputation: 153460

Let's try this:

df1 = df.groupby('id')['vote'].apply(lambda x: pd.DataFrame(x.tolist())).unstack()
df1.columns = df1.columns.droplevel().values + 1
df1.add_prefix('vote_').reset_index()

Output:

         id vote_1 vote_2 vote_3 vote_4
0  6690d79f    Pos    Neg    Pos    Neu
1  7690d79f    Pos    Neg    Pos    Neu

Upvotes: 2

Related Questions