looking 4 py code: line up columns

[Skaperen]

i am looking for a program to line up (white-space by default, otherwise specified string) separated columns, given in a list of strings (lines) for a function/module implementation, or in a file or stdin for a command implementation.  the columns should be left justified unless all rows of that column are decimal numbers in which case they are right justified. a more advanced version could also line up CIDR addresses so the '/' characters in them line up.

[sparkz_alot]

This tutorial gives some excellent examples of formatting.  https://docs.python.org/3/tutorial/input...formatting

[Skaperen]

one of the complications of a program like this is that it needs to figure out the width of each column, for as many columns as are present, across all rows.  it also needs verify the datatype of each column across all rows.  and it gets more complex when working with different justifications on adjacent columns.

[ichabod801]

I've done this before, I just can't remember where. It's not that hard. Loop through the columns, determining the maximum width of the string representations. Construct a format string (with {}'s) using the width information plus 1 or 2 spaces for a gutter. Loop through the rows, passing the row to the format method with * notation. Join the resulting strings with '\n'. You just have the added wrinkle of detecting numeric columns and putting that into the format specification.

[wavic]

How about Pandas module? print(pandas.DataFrame.from_records(data))

[Skaperen]

i mean text data that may be numbers or may be words and names, more often the latter.  i want to line them up for printing.  here is a list states i have been in and their capitals, lined up, sorted by state and boxed:

Output:
+---<STDIN>-------------------+
| Arkansas       Little_Rock  |
| Delaware       Dover        |
| Illinois       Springfield  |
| Indiana        Indianapolis |
| Kentucky       Frankfort    |
| New_Jersey     Trenton      |
| New_York       Albany       |
| North_Carolina Raleigh      |
| Ohilo          Columbus     |
| Pennsylvania   Harrisburg   |
| South_Carolina Columbia     |
| Texas          Austin       |
| Virginia       Richmond     |
| West_Virginia  Charleston   |
+-----------------------------+

[ichabod801]

Here's my attempt to recreate my code just after waking up:

states = [['Arkansas', 'Little_Rock'], ['Delaware', 'Dover'], ['Illinois', 'Springfield'], ['Indiana', 'Indianapolis'], 
   ['Kentucky', 'Frankfort'], ['New_Jersey', 'Trenton'], ['New_York', 'Albany'], ['North_Carolina', 'Raleigh'], 
   ['Ohilo', 'Columbus'], ['Pennsylvania', 'Harrisburg'], ['South_Carolina', 'Columbia'], ['Texas', 'Austin'], 
   ['Virginia', 'Richmond'], ['West_Virginia', 'Charleston']]

max_widths = []
for column_index in range(len(states[0])):
   column = [row[column_index] for row in states]
   max_widths.append(len(max([str(datum) for datum in column], key = len)))
specs = ['{{:<{}}}'.format(width) for width in max_widths]
format_spec = ' '.join(specs)
lines = [format_spec.format(*row) for row in states]
print('\n'.join(lines))

Here's the output:

Output:
Arkansas       Little_Rock 
Delaware       Dover       
Illinois       Springfield 
Indiana        Indianapolis
Kentucky       Frankfort   
New_Jersey     Trenton     
New_York       Albany      
North_Carolina Raleigh     
Ohilo          Columbus    
Pennsylvania   Harrisburg  
South_Carolina Columbia    
Texas          Austin      
Virginia       Richmond    
West_Virginia  Charleston  

You would just need to vary the justification indicator in the format specification based on the type of the column.

[snippsat]

String formatting as advised.
Example:

data = '''\
Arkansas,Little_Rock
Delaware,Dover
Illinois,Springfield
Indiana,Indianapolis
Kentucky,Frankfort'''

data_list = [line.split(',') for line in data.split('\n')]
header = ['State', 'Capital']

sf = "{:<10s} {:>5s}"
print(sf.format(*header))
print('-'*23)
for sublist in data_list:
   print(sf.format(*sublist))
print('-'*23)

Output:
State      Capital
-----------------------
Arkansas   Little_Rock
Delaware   Dover
Illinois   Springfield
Indiana    Indianapolis
Kentucky   Frankfort
-----------------------

[Skaperen]

I've done this before, I just can't remember where. It's not that hard. Loop through the columns, determining the maximum width of the string representations. Construct a format string (with {}'s) using the width information plus 1 or 2 spaces for a gutter. Loop through the rows, passing the row to the format method with * notation. Join the resulting strings with '\n'. You just have the added wrinkle of detecting numeric columns and putting that into the format specification.

one thing i was looking for was to see how intermixing of column types was handled.  if left justified column is to the the left of a right justified column it leave open the option to squeeze them tighter.  would you (dare to) do that?  that's code i'd like to see.

FYI, i did this decades ago in 370 assembler ... with the justification mix.