pandas - convert Series to columns

import pandas as pd

def test1():

    df = pd.DataFrame({
        'col1': [1, 2, 3],
        'colX': [
            pd.Series([1, 2, 3, None, 0, '']),
            pd.Series([4, 5, 6, None, 0, '']),
            pd.Series([7, 8, 9, None, 0, '']),
        ]
    })

    print('\n--- info ---\n')
    df.info()
    print('\n--- head ---\n')
    print(df.head())
    print('\n--- series dtype ---\n')
    print(df['colX'][0].dtype)
    print('---')

    df = df.astype({
        "col1":  float,
        #"colX":  float,
    })

    print('\n--- info ---\n')
    df.info()
    print('\n--- head ---\n')
    print(df.head())
    print('\n--- series dtype ---\n')
    print(df['colX'][0].dtype)
    print('---')

    print(round( df.isnull().sum(axis=0)/len(df)*100, 2 ))

    print(round( 100 - df.count()/len(df)*100, 2 ))

def test2():

    df = pd.DataFrame({
        'col1': [1, 2, 3],
        'colX': [
            pd.Series([1, 2, 3, None, 0, '']),
            pd.Series([4, 5, 6, None, 0, '']),
            pd.Series([7, 8, 9, None, 0, '']),
        ]
    })

    # split to columns (and convert to float at the same time)
    #df['X_0'] = df['colX'].str[0]
    #df['X_1'] = df['colX'].str[1]
    #df['X_2'] = df['colX'].str[2]
    #df['X_3'] = df['colX'].str[3]
    df['X_0'] = df['colX'].apply(lambda x:x[0])#.astype(float)
    df['X_1'] = df['colX'].apply(lambda x:x[1])#.astype(float)
    df['X_2'] = df['colX'].apply(lambda x:x[2])#.astype(float)
    df['X_3'] = df['colX'].apply(lambda x:x[3])#.astype(float)
    df['X_4'] = df['colX'].apply(lambda x:x[4])#.astype(float)
    df['X_5'] = df['colX'].apply(lambda x:x[5])#.astype(float)
    # remove column with series
    df.drop(columns='colX', inplace=True)

    #df.info()
    #print(df.head())

    # convert other columns (no need to convert `X_`)
    df = df.astype({
        "col1":  float,
        #"X_0":  float,
        #"X_1":  float,
        #"X_2":  float,
        #"X_3":  float,
        #"X_4":  float,
        #"X_5":  float,
    })

    #df.info()
    #print(df.head())

    print(round( df.isnull().sum(axis=0)/len(df)*100, 2 ))

    print(round( 100 - df.count()/len(df)*100, 2 ))

def test3():

    import pandas as pd

    df = pd.DataFrame({
        'col1': [1, 2, 3],
        'colX': [
            pd.Series([1, 2, 3, None, 0, '']),
            pd.Series([4, 5, 6, None, 0, '']),
            pd.Series([7, 8, 9, None, 0, '']),
        ]
    })

    #print(df['colX'][0].dtype)
    #print(df.head())

    #print(round((df.isnull().sum(axis=0)/len(df.index))*100, 2))

    # convert to float values in Series
    #df['colX'] = df['colX'].apply(lambda x: x.astype(float))

    # convert Series to columns with prefix `X_` (need to apply pd.Series even if there are series)
    df = df.assign( **(df['colX'].apply(pd.Series).add_prefix('X_')) )

    # remove column with series
    df.drop(columns='colX', inplace=True)

    #df.info()
    #print(df.head())

    # convert other columns (no need to convert `X_`)
    df = df.astype({
        "col1":  float,
        #"X_0":  float,
        #"X_1":  float,
        #"X_2":  float,
        #"X_3":  float,
        #"X_4":  float,
        #"X_5":  float,
    })

    #df.info()
    #print(df.head())

    print(round( df.isnull().sum(axis=0)/len(df)*100, 2 ))

    print(round( 100 - df.count()/len(df)*100, 2 ))

test1()
test2()
test3()