Join two floats with trailing 0 together

Question

I wrote a function:

def main_table(data, gby_lst, col):
    df = data.groupby(gby_lst)[col].describe()
    df = df.reset_index()
    for i in ['25%', '50%', '75%', 'std', 'min', 'max', 'mean']:
        df[i] = df[i].apply(lambda x: float("{:.2f}".format(x)))
    df['Mean ± SD'] = (df[['mean', 'std']]
                       .apply(lambda row: ' ± '.join(row.values.astype(str)), axis=1)
                      )
    df['Median (IQR)'] = (df['50%'].astype(str) + ' (' + df[['25%', '75%']].apply(lambda row: ' - '.join(row.values.astype(str)),
                                                axis=1) + ')'
                         )
    df['Range'] = (df[['min', 'max']]
                   .apply(lambda row: ' - '.join(row.values.astype(str)), axis=1)
                  )


    summary_list = gby_lst + ['Mean ± SD', 'Median (IQR)', 'Range']
    return df.loc[:, summary_list]

But this will not include the ending 0s. For example, I want 3.40 ± 5.55 , this function currently gives me: 3.4 ± 5.55.

How can I fix it?

Answer 1

Change the line from:

df[i] = df[i].apply(lambda x: float("{:.2f}".format(x)))

To this:

df[i] = df[i].apply(lambda x: "{:.2f}".format(x))

Python's float() function will truncate 0s by default to save space, which it is doing when converting from a string in this example. When using just the formatter {:.2f} we are explicitly formatting the string to have exactly 2 decimal places. In the first line, the formatter specifications are being overwritten after we cast it as float.