[FIXED] How to format a dataframe column wise in pandas?

Issue

I used to format the dataframe and output a file by:

tta = pandas.DataFrame(
    {'name': ['xx', 'y y', 'z'], 'alter': ['AA', 'BB', 'CCC']})
str1 = tta.to_string(formatters={'name': '{:05s}'.format,
                                 'alter': '{:08s}'.format
                                 }, header=False, index=False, justify='right', col_space=0)

However, the to_string function adds an extra space to my output string like the str1 in the script:

'xx000 AA000000\ny y00 BB000000\nz0000 CCC00000'

It seems it is hard to remove the extra white space. (See: Remove the automatic two spaces between columns that Pandas DataFrame.to_string inserts)

Thus is there any elegant way to format the data frame based on the column and let me get results like this:

'xx000AA000000\ny y00BB000000\nz0000CCC00000'

Solution

I don't think there is a straightforward way to achieve this, the space separators seem hardcoded in pandas' source, I didn't find an obvious way to change them. A simple hack however might be to add a custom placeholder (e.g. -*-) in your format to easily be able to remove it together with the following space afterwards:

PLACEHOLDER = '-*-'
str1 = (tta.to_string(formatters={'name': ('{:05s}'+PLACEHOLDER).format,
                                  'alter': '{:08s}'.format
                                  },
                      header=False, index=False, justify='right', col_space=0)
           .replace(f'{PLACEHOLDER} ', '')
       )

Output:

'xx000AA000000\ny y00BB000000\nz0000CCC00000'

Answered By - mozway