python reshape every nth column - pandas

Question

I have just started with python and need some help. I have a dataframe which looks like "Input Data", What I want is stack by every nth column. In other words, I want a dataframe where every nth Column is appended below to first m rows
id
city
Col 1
Col 2
Col 3
Col 4
Col 5
Col 6
Col 7
Col 8
Col 9
Col 10
1
1
51
155
255
355
455
666
777
955
55
553
2
0
52
155
255
355
455
666
777
595
55
553
3
NAN
53
155
255
355
455
666
777
559
55
535
4
1
54
155
255
355
545
666
777
559
55
535
5
7
55
155
255
355
455
666
777
955
55
535
Required Output
id
city
Col 1
Col 2
Col 3
Col 4
Col 5
1
1
51
155
255
355
455
2
0
52
155
255
355
455
3
NAN
53
155
255
355
455
4
1
54
155
255
355
545
5
7
55
155
255
355
455
1
1
666
777
955
55
553
2
0
666
777
595
55
553
3
NAN
666
777
559
55
535
4
1
666
777
559
55
535
5
7
666
777
955
55
535
I am trying to do something opposite of this
Input & required Output

Answer 1

In [74]: column_list = [df.columns[k:k+5] for k in range(2, len(df.columns), 5)]
In [75]: column_list
Out[75]:
[Index(['Col 1', 'Col 2', 'Col 3', 'Col 4', 'Col 5'], dtype='object'),
Index(['Col 6', 'Col 7', 'Col 8', 'Col 9', 'Col 10'], dtype='object')]
In [76]: dfs = [df[['id', 'city'] + columns.tolist()].rename(columns=dict(zip(columns, range(5)))) for columns in column_list]
In [77]: dfs
Out[77]:
[ id city 0 1 2 3 4
0 1 1.0 51 155 255 355 455
1 2 0.0 52 155 255 355 455
2 3 NaN 53 155 255 355 455
3 4 1.0 54 155 255 355 545
4 5 7.0 55 155 255 355 455,
id city 0 1 2 3 4
0 1 1.0 666 777 955 55 553
1 2 0.0 666 777 595 55 553
2 3 NaN 666 777 559 55 535
3 4 1.0 666 777 559 55 535
4 5 7.0 666 777 955 55 535]
In [78]: pd.concat(dfs, ignore_index=True)
Out[78]:
id city 0 1 2 3 4
0 1 1.0 51 155 255 355 455
1 2 0.0 52 155 255 355 455
2 3 NaN 53 155 255 355 455
3 4 1.0 54 155 255 355 545
4 5 7.0 55 155 255 355 455
5 1 1.0 666 777 955 55 553
6 2 0.0 666 777 595 55 553
7 3 NaN 666 777 559 55 535
8 4 1.0 666 777 559 55 535
9 5 7.0 666 777 955 55 535
To explain :
First generate the required columns for each slice
pd.concat requires the column names of all the dataframes in the list to be the same, hence the renames in rename(columns=dict(zip(columns, range(5)))). We are just renaming the sliced columns to 0,1,2,3,4
Last step is to concat everything.
EDIT
Based on the comments by OP:
Sorry #Asish M. but how to add a column for dataset number in each dataset of dfs, eg- here we split our dataset into 2, so I need one column which says for first 1 to 5 ids - 'first' (or 1), then again for another 1 to 5 ids - 'second' (or 2) in the output. I hope it's making scenes
dfs = [df[['id', 'city'] + columns.tolist()].assign(split_group=idx).rename(columns=dict(zip(columns, range(5)))) for idx, columns in enumerate(column_list)]
df.assign(split_group=idx) creates a column 'split_group' with value = idx. You get the idx from enumerating the column_list