Sort a CSV row in bash

问题

The output from a command is sent as a CSV list of UUIDs. The UUIDs are not sorted however, so it's very difficult to tell if a line is unique. I would like to sort each line by the value between the commas, and then uniq the lines.

I know I could hack something up with awk, but I was hoping for a cleaner/more elegant one-liner. Any ideas?

EDIT

Here is some sample data:

9166e19c-4794-467e-baad-3f8c2f2656cb,f5553f54-589b-4afd-a8e0-2239b23dc138,ee721e70-a7e2-4da2-a2b0-22bec3432c3d,7e17bf09-e56b-428e-94c9-a7dc50991e00,360b7de7-d7e5-455a-8eb8-0bd856c705ed
9166e19c-4794-467e-baad-3f8c2f2656cb,360b7de7-d7e5-455a-8eb8-0bd856c705ed,7e17bf09-e56b-428e-94c9-a7dc50991e00,f5553f54-589b-4afd-a8e0-2239b23dc138,ee721e70-a7e2-4da2-a2b0-22bec3432c3d
9166e19c-4794-467e-baad-3f8c2f2656cb,ee721e70-a7e2-4da2-a2b0-22bec3432c3d,7e17bf09-e56b-428e-94c9-a7dc50991e00,f5553f54-589b-4afd-a8e0-2239b23dc138,360b7de7-d7e5-455a-8eb8-0bd856c705ed
9166e19c-4794-467e-baad-3f8c2f2656cb,ee721e70-a7e2-4da2-a2b0-22bec3432c3d,360b7de7-d7e5-455a-8eb8-0bd856c705ed,7e17bf09-e56b-428e-94c9-a7dc50991e00,f5553f54-589b-4afd-a8e0-2239b23dc138
9166e19c-4794-467e-baad-3f8c2f2656cb,f5553f54-589b-4afd-a8e0-2239b23dc138,360b7de7-d7e5-455a-8eb8-0bd856c705ed,ee721e70-a7e2-4da2-a2b0-22bec3432c3d,7e17bf09-e56b-428e-94c9-a7dc50991e00
9166e19c-4794-467e-baad-3f8c2f2656cb,360b7de7-d7e5-455a-8eb8-0bd856c705ed,ee721e70-a7e2-4da2-a2b0-22bec3432c3d,f5553f54-589b-4afd-a8e0-2239b23dc138,7e17bf09-e56b-428e-94c9-a7dc50991e00
9166e19c-4794-467e-baad-3f8c2f2656cb,7e17bf09-e56b-428e-94c9-a7dc50991e00,f5553f54-589b-4afd-a8e0-2239b23dc138,360b7de7-d7e5-455a-8eb8-0bd856c705ed,ee721e70-a7e2-4da2-a2b0-22bec3432c3d
9166e19c-4794-467e-baad-3f8c2f2656cb,360b7de7-d7e5-455a-8eb8-0bd856c705ed,ee721e70-a7e2-4da2-a2b0-22bec3432c3d,f5553f54-589b-4afd-a8e0-2239b23dc138,7e17bf09-e56b-428e-94c9-a7dc50991e00
9166e19c-4794-467e-baad-3f8c2f2656cb,f5553f54-589b-4afd-a8e0-2239b23dc138,360b7de7-d7e5-455a-8eb8-0bd856c705ed,7e17bf09-e56b-428e-94c9-a7dc50991e00,ee721e70-a7e2-4da2-a2b0-22bec3432c3d
9166e19c-4794-467e-baad-3f8c2f2656cb,7e17bf09-e56b-428e-94c9-a7dc50991e00,360b7de7-d7e5-455a-8eb8-0bd856c705ed,f5553f54-589b-4afd-a8e0-2239b23dc138,ee721e70-a7e2-4da2-a2b0-22bec3432c3d
9166e19c-4794-467e-baad-3f8c2f2656cb,7e17bf09-e56b-428e-94c9-a7dc50991e00,ee721e70-a7e2-4da2-a2b0-22bec3432c3d,360b7de7-d7e5-455a-8eb8-0bd856c705ed,f5553f54-589b-4afd-a8e0-2239b23dc138
9166e19c-4794-467e-baad-3f8c2f2656cb,ee721e70-a7e2-4da2-a2b0-22bec3432c3d,7e17bf09-e56b-428e-94c9-a7dc50991e00,360b7de7-d7e5-455a-8eb8-0bd856c705ed,f5553f54-589b-4afd-a8e0-2239b23dc138
9166e19c-4794-467e-baad-3f8c2f2656cb,f5553f54-589b-4afd-a8e0-2239b23dc138,360b7de7-d7e5-455a-8eb8-0bd856c705ed,7e17bf09-e56b-428e-94c9-a7dc50991e00,ee721e70-a7e2-4da2-a2b0-22bec3432c3d
9166e19c-4794-467e-baad-3f8c2f2656cb,360b7de7-d7e5-455a-8eb8-0bd856c705ed,7e17bf09-e56b-428e-94c9-a7dc50991e00,f5553f54-589b-4afd-a8e0-2239b23dc138,ee721e70-a7e2-4da2-a2b0-22bec3432c3d

Cheers.

回答1:

With Perl and uniq you can do it like this:

perl -F, -lane '@A=sort @F; print join(",",@A)' input_file | uniq

EDIT:

@A is actually unneccessary, this works as well and is shorter and faster.

perl -F, -lane 'print join(",",sort @F)' input_file | uniq

Options used:

• -e = may be used to enter one line of program (one liner mode)
• -a = turn on autosplit mode, puts values from split on @F array
• -F, = set split delimiter to comma
• -n = causes perl to assume while(<>){...} loop around your program
• -l = enables automatic line-ending processing

In this line @F is a special array containing UUIDs that were split. It's sorted and copied to @A array. Then @A is printed with values being joined by ,. Output from this command is piped to uniq so you get unique lines only.

Output:

360b7de7-d7e5-455a-8eb8-0bd856c705ed,7e17bf09-e56b-428e-94c9-a7dc50991e00,9166e19c-4794-467e-baad-3f8c2f2656cb,ee721e70-a7e2-4da2-a2b0-22bec3432c3d,f5553f54-589b-4afd-a8e0-2239b23dc138

回答2:

using shell

while read i 
do
  echo $i | tr "," "\n" | sort | tr "\n" "," | sed "s/,$//"
  echo 
done < file |sort -u

来源：https://stackoverflow.com/questions/21468143/sort-a-csv-row-in-bash