Using Python's Format Specification Mini-Language to specify or automatically determine column widths

Question

I have the following example in Python to print a table:

job_IDs = ['13453', '123', '563456'];
memory_used = [30, 150.54, 20.6];
memory_units = ['MB', 'GB', 'MB'];

for i in range(len(job_IDs)):
    print "Job {item:15} {value[0]:>6}.{value[1]:<6} {units:3}".format(
    item=job_IDs[i]+':', value=memory_used[i].split('.') if '.' in memory_used[i] 
    else  (memory_used[i], '0'), units=memory_units[i])

which results in:

Job 13453:                 30          MB
Job 123:                  150.54       MB
Job 563456:                20.6        GB

The vertical alignment is correct for what I need, but the horizontal padding is manually hard-coded in the code, where it says:

"Job {item:15} {value[0]:>6}.{value[1]:<6} {units:3}"

(the "padding parameters" above are 15, 6, 6, 3)

I would like to:

• Ideally, have the padding values determined automatically (i.e. computed from the maximum width of each field) instead of having to hard-code them in the statement above.

  In the example above, it would be great to have a snippet that uses the width of the widest entries in job_IDs, memory_used, and memory_units and outputs, leaving one space between columns:

  Job 13453:   30    MB
  Job 123:    150.54 MB
  Job 563456:  20.6  GB
  

• Alternatively, I would also interested in a solution that allows me to specify such paddings in variables (in case I want to control the width given to each column).

I am interested in solutions for both problems: with the padding computed automatically, and with the padding specified with variables.

By the way, This question was inspired by this other question: Using Python's Format Specification Mini-Language to align floats

Answer 1

You can nest the arguments to change the integer width like so:

job_IDs = ['13453', '123', '563456'];
memory_used = ['30', '150.54', '20.6'];
memory_units = ['MB', 'GB', 'MB'];

w1=15
w2=6
w3=6
w4=3
for i in range(len(job_IDs)):
    v=memory_used[i].split('.') if '.' in memory_used[i] else  (memory_used[i], '0')

    print "Job {item:{w1}} {value[0]:>{w2}}.{value[1]:<{w3}} {units:{w4}}".format(
       item=job_IDs[i]+':', w1=w1, w2=w2,w3=w3,w4=w4, value=v, units=memory_units[i])

So to determine the field width to use, just loop through one time before you print to determine the widest width to use for each applicable embedded width variable.

Like so:

job_IDs = ['13453', '123', '563456'];
memory_used = ['30', '150.54', '20.6'];
memory_units = ['MB', 'GB', 'MB'];

w1=1
w2=1
w3=1
w4=1
for i in range(len(job_IDs)):
    v=memory_used[i].split('.') if '.' in memory_used[i] else  (memory_used[i], '0')
    item=job_IDs[i]+':'
    units=memory_units[i]
    if len(item)>w1: w1=len(item)
    if len(v[0])>w2: w2=len(v[0])
    if len(v[1])>w3: w3=len(v[1])
    if len(units)>w4: w3=len(units)

for i in range(len(job_IDs)):
    v=memory_used[i].split('.') if '.' in memory_used[i] else  (memory_used[i], '0')
    item=job_IDs[i]+':'
    units=memory_units[i]    
    print "Job {item:{w1}} {value[0]:>{w2}}.{value[1]:<{w3}} {units:{w4}}".format(
      item=item,w1=w1, w2=w2,w3=w3,w4=w4, value=v, units=units)

Output:

Job 13453:   30.0  MB
Job 123:    150.54 GB
Job 563456:  20.6  MB