Use values of specific rows as columns names

Question

I have multiple dataframes that looks like:

        0  1  2  3  4  5
Groupe  1  2  3  4  5  6
1       3  6  8  6  8  9
2       5  5  2  1  1   0.1
3       7  2  8  1.9 11 2.1
4       1  2  1  1.1  2  1

So the columns names are [0, 1, 2, 3, 4, 5] but I’d like to put the values of Groupe row as columns name, because it’s interesting in the context that I analyze the data.
Remarque: the values of Groupe row could change, so it’s not necessarily 1, 2, 3. so the solution could not be just as trivial as to add +1 to actual column names.

the desired output for the dataframe given above is:

       1  2  3  4  5  6
1       3  6  8  6  8  9
2       5  5  2  1  1   0.1
3       7  2  8  1.9 11 2.1
4       1  2  1  1.1  2  1

So how can I change the columns names for values of some row ?

Asked By: Amanda

||

Source

Answer 1

If row Groupe is first row of df, you can first set column names by first row selected by iloc, then convert it to int by astype, remove column name and last remove first row of DataFrame with [1:]:

df.columns = df.iloc[0,:].astype(int)
df.columns.name= None
print (df[1:])
   1  2  3    4   5    6
1  3  6  8  6.0   8  9.0
2  5  5  2  1.0   1  0.1
3  7  2  8  1.9  11  2.1
4  1  2  1  1.1   2  1.0

Another solution with loc and drop, row Groupe in df can be anywhere, not only first:

df.columns = df.loc['Groupe',:].astype(int)
df.columns.name= None
df = df.drop('Groupe')
print (df)
   1  2  3    4   5    6
1  3  6  8  6.0   8  9.0
2  5  5  2  1.0   1  0.1
3  7  2  8  1.9  11  2.1
4  1  2  1  1.1   2  1.0

Answered By: jezrael

Use values of specific rows as columns names

Question:

Answers: