Split one single column with periodic repeated headers into two columns (Python)-CodePudding

I have one single column that looks like this

None header
time(s)
0.2055
0.2302
position(m)
0.4164
0.4128
time(s)
0.2055
0.2302
position(m)
0.4164
0.4128

And I want split this column into columns like this:

time(s)	position(m)
...	...

CodePudding user response：

You can use numpy reshaping:

N = len(df)//4

m = np.tile(np.repeat([False,True], [1,N-1]), len(df)//N)
# array([False,  True,  True, False,  True,  True, False,  True,  True,
#        False,  True,  True])

df2 = pd.DataFrame(df.iloc[m, 0]
                     .to_numpy()
                     .reshape((2,-1,N-1))
                     .swapaxes(1,2)
                     .reshape(-1,2),
                   columns=df.iloc[[0,N],0].to_list()
                   )
print(df2)

Output:

  time(s) position(m)
0  0.2055      0.4164
1  0.2302      0.4128
2  0.2055      0.4164
3  0.2302      0.4128

CodePudding user response：

Use:

df['g'] = df.index//3%2
temp = df.groupby('g')['header'].apply(list).to_frame().T.explode([0,1]).reset_index(drop=True)
temp.columns = temp.iloc[0]
temp[~temp['time(s)'].str.contains('[a-z]')]

Output:

    time(s) position(m)
1   0.2055  0.4164
2   0.2302  0.4128
4   0.2055  0.4164
5   0.2302  0.4128