Split list in a dataframe in multiple columns using one as the header-CodePudding

Suppose I got a pandas dataframe with two columns containing a list (length >= 1) at the end. The first one ("mode") has a variable which should be appended to the desired header name, the second column ("res") contains the data:

>>> df = pd.DataFrame([ 
    { 'c1': 850, 'c2': 'Ex', 'c3': 300.0, 'c4': 250, 'mode': [0, 1], 'res': [1.525, 1.321] },
    { 'c1': 850, 'c2': 'Ex', 'c3': 300.0, 'c4': 250, 'mode': [0, 1], 'res': [1.526, 1.311] }
])

with the result

    c1  c2     c3   c4    mode             res
0  850  Ex  300.0  250  [0, 1]  [1.525, 1.321]
1  850  Ex  300.0  250  [0, 1]  [1.526, 1.311]

Is there a better way to split the dataframe df to get this desired result

    c1  c2  c3      c4   res_mode_0  res_mode_1
0  850  Ex  300.0   250       1.525       1.321
1  850  Ex  310.0   250       1.526       1.311

than using loops?

CodePudding user response：

You can try the following code. Advantage is it works regardless of the number of elements in the list.

df = pd.DataFrame([ 
    { 'c1': 850, 'c2': 'Ex', 'c3': 300.0, 'c4': 250, 'mode': [0, 1], 'res': [1.525, 1.321] },
    { 'c1': 850, 'c2': 'Ex', 'c3': 300.0, 'c4': 250, 'mode': [0, 1], 'res': [1.526, 1.311] }
])

split_df = pd.DataFrame(df["res"].tolist()).add_prefix("res_mode_")

df = pd.concat([df, split_df], axis=1).drop(["mode", "res"], axis=1)

Output:

df

    c1      c2  c3      c4      res_mode_0  res_mode_1
0   850     Ex  300.0   250     1.525       1.321
1   850     Ex  300.0   250     1.526       1.311

CodePudding user response：

The most efficient way to do it;

pd.concat([pd.DataFrame(df.pop('your_column').values.tolist()), df], axis=1)

Unfortunately, you will have to use this on each columns you need to expand.