Split 1-column dataframe rows into new colums-CodePudding

I have a dataframe with x rows and would like to split it into x rows into column in a new dataframe.

I have found this example (30 rows in the x dataframe), where it works if I split it to something where row x col = 30 (below 10 rows en each columns => 3 col)

k = pd.concat([pd.Series(j, name='y'   str(i), index=range(0,10)) for i,j in enumerate(np.array_split(x['TEST'].values, 3))], axis=1)

But if I would like e.g. 8 rows in columns which mean 4 columns, the last column would not be of the size 8 and I get this error

ValueError: Length of values (7) does not match length of index (8)

So how can I split a column into x chunks and then automatic get y columns in new dataframe?

CodePudding user response：

Instead of specifying to number of chunks you can specify the indices where to split:

x = pd.DataFrame({'TEST': range(30)})
n = 8
pd.concat([pd.Series(j, name='y'   str(i)) for i,j in enumerate(np.split(x['TEST'].to_numpy(), range(n, len(x['TEST']), n)))], axis=1)

Result:

   y0  y1  y2    y3
0   0   8  16  24.0
1   1   9  17  25.0
2   2  10  18  26.0
3   3  11  19  27.0
4   4  12  20  28.0
5   5  13  21  29.0
6   6  14  22   NaN
7   7  15  23   NaN