How to left align column values in pandas `to_string()`?

Question

I want to save a pandas dataframe to a file with to_string(), but want to left align the column values. With to_string(justify=left), only column labels are left aligned.

For example with

pd.DataFrame({'col1': [' 123 ', ' 1234'], 'col2': ['1', '444441234']}).to_string(index=False)

I get the following result:

I want to get rid of the whitespaces in the first row by left aligning the column values.

Answer 1

The to_string methods provides support for per column formatters. It allows you to use specific formats for all of some columns. A rather simple way is to create a format and then apply it with a lambda. The only picky part is that to use left formatting, you will have to know the width of the column.

For your provided data, you could left align everything with:

df = pd.DataFrame({'col1': ['   123 ', ' 1234'], 'col2': ['1', '444441234']})
widths = [4, 9]
formats = ['{' + f':<{i}' + '}' for i in widths]
print(df.to_string(index=None, col_space=widths, formatters=
                   [(lambda x: fmt.format(x)) for fmt in formats],
                   justify='left'))

to get:

col1       col2      
    123     1        
  1234      444441234

You could also left align only some columns by using a dict for the formatters parameter:

print(df.to_string(index=None, formatters=
                   {'col2': (lambda x: '{:<9}'.format(x))},
                   justify='left'))

gives:

col1     col2      
    123   1        
    1234  444441234