Sort script output but keep lines together

Works for files, which contains two lines, only.

sed -n 'N; s/\n//p' *.conf | sort | sed -n 's/\(Package\)/\n\1/p'

Explanation:

1. sed -n 'N; s/\n//p' *.conf - removes newline between two lines, that is, it joins two contiguous lines in to the one.
2. sort - sorts long lines, which was made in the previous step.
3. sed -n 's/\(Package\)/\n\1/p' - returns newline back to its place.

Output:

Order: 0 Package: example Order: 1 Package: file Order: 2 Package: another 

A naive Unix aproach:

awk '/Order|Package/' conf/*.conf | # get your usual output paste - - | # convert two lines to one sort -n -k2,2 | # sort numerically awk '{print $1,$2;print $3,$4}' # re-print in two columns 

Python script

#!/usr/bin/env python from __future__ import print_function import sys def get_order_package(filename): order=None package=None with open(filename) as f: for line in f: if line.startswith('Order:'): order=int(line.strip().split()[1]) if line.startswith('Package:'): package=line.strip().replace("Package: ",'') return (order,package) order_list=[] for argument in sys.argv[1:]: order_list.append(get_order_package(argument)) order_list.sort(key=lambda x: x[0]) for i in order_list: print('Order:',i[0]) print('Package:',i[1]) 

This script defines get_order_package() function which parses the file, and returns tuple in format (integer,package_string). We pass all files as command-line arguments and iterate over them, executing the function explained above for each one. That builds up a list, we sort it, and iterate again, printing each respective tuple parts from the new sorted list.

Test run

$ cat example_1.conf Order: 0 Package: example_1 $ cat example_2.conf Order: 2 Package: example_2 $ cat example_3.conf Order: 1 Package: example_3 $ ./parse_conf.py *.conf Order: 0 Package: example_1 Order: 1 Package: example_3 Order: 2 Package: example_2 

This works too:

grep -e ^Order -e ^Package `grep ^Order *.conf | awk -F: '{printf("%s:%s\n",$NF,$0);}' |sort| awk -F: '{print $2}'` 

Explanation (below multi-line format is only for demonstration, it does not work in bash because the resultant single-line command cannot handle the comments):

$ grep -e ^Order -e ^Package ` \ # grep for lines starting with "Order" or "Package"; start nested command > grep ^Order *.conf | \ # grep for lines starting with "Order" in files *.conf > awk -F: '{printf("%s:%s\n",$NF,$0);}' | \ # prefixing order number with a separating colon > sort | \ # sort by order number > awk -F: '{print $2}' \ # extract file name alone ; now filenames are ordered by the order number contained within them > ` # end nested command