CODEWITHSUNDEEP
pythonpandas
user88484
user88484

Reputation: 1557

move rows to the first row (as new columns) pandas

I'm kind of new to pandas and I'm trying to perform some table manipulation. I'm not exactly sure how the action I'm looking for is called.

I have a pandas dataframe (the actual one is larger, but these are the cols I need to convert:

   pr10      pr50      pr90
25.6214   30.2753    36.634
1.94509   3.84798   6.55111
      0         0   0.00835
      0         0  0.007279

And I want to convert it to:

pr10_0  pr50_0  pr90_0  pr10_1  pr50_1  pr90_1  pr10_2  pr50_2  pr90_2  pr10_3  pr50_3  pr90_3  pr10_4  pr50_4  pr90_4
25.6214 30.2753 36.634  1.94509 3.84798 6.55111 0   0   0.00835 0   0   0.007279    0.003505    0.007009    0.014764

How can I accomplish that?

Upvotes: 2

Views: 1443

Answers (2)

DJK
DJK

Reputation: 9264

You could recreate the dataframe:

Create new columns from current dataframe

cols = [name+'_{}'.format(num // len(df.columns)) for num,name in enumerate(df.columns.tolist()*len(df))]

Create a new dataframe by flattening all the values from the original dataframe

pd.DataFrame(df.values.reshape(1,-1),columns=cols)

   pr10_0   pr50_0  pr90_0   pr10_1   pr50_1   pr90_1  pr10_2  pr50_2  \
0  25.6214  30.2753  36.634  1.94509  3.84798  6.55111     0.0     0.0   
    pr90_2  pr10_3  pr50_3    pr90_3  
0  0.00835     0.0     0.0  0.007279

Upvotes: 0

BENY
BENY

Reputation: 323226

Using stack and join the multiple index 

df1=df.stack().to_frame(0).T

df1.columns=df1.columns.map('{0[1]}_{0[0]}'.format) 
df1
Out[445]: 
    pr10_0   pr50_0  pr90_0   pr10_1   pr50_1   pr90_1  pr10_2  pr50_2  \
0  25.6214  30.2753  36.634  1.94509  3.84798  6.55111     0.0     0.0   
    pr90_2  pr10_3  pr50_3    pr90_3  
0  0.00835     0.0     0.0  0.007279

Upvotes: 1

Related Questions