Batch rename a lot of file using a number find in filename

Question

I have hundred of file name with a number tag in filename like this:

AAAA #12 SSSS.TXT
BBB #231 CDF.TXT
CDFSDAAAA #1 AAAASS.TXT

and i want to rename so:

#012 - AAAA #12 SSSS.TXT
#231 - BBB #231 CDF.TXT
#001 - CDFSDAAAA #1 AAAASS.TXT

How I can do it?

Thanks

Is the "number tag" always the 2nd space-separated portion of the filename, and always following a #-mark? — Jeff Schaller, Feb 01 '19 at 18:37
no, I am unlucky... maybe the second, the third or more but is ever in the pattern #000 one #, 1 or more number and a space... — Ancucchi, Feb 01 '19 at 18:38
yes
digit digit digit space

the digit usually are from 1 to 4 — Ancucchi, Feb 01 '19 at 18:41
Is it possible that # will appear more than once in the filename? — jesse_b, Feb 01 '19 at 18:42
(probably) my last question: do you have to use find here, because they're all in various subdirectories, or are they all in the same directory? — Jeff Schaller, Feb 01 '19 at 18:42
they are all in the same folder and the # appear one time in every filename... — Ancucchi, Feb 01 '19 at 18:43

score 3 · Accepted Answer · answered Feb 01 '19 at 18:52

In the bash shell,

for file in *.TXT
do 
  [[ $file =~ \#([[:digit:]]+) ]] && mv -- "$file" "#$(printf '%03d' "${BASH_REMATCH[1]}") - ${file}"
done

This loops over all of the files in the current directory that end in .TXT and compares them with bash's =~ conditional expression operator. It compares the incoming filename to the regular expression on the right. The regex looks for a hash mark (escaped, so it's not a comment), followed by some (captured in parenthesis) digits. The + is greedy, so it will pick up as much of the number as it can (1 digit, 4 digits, etc), ending at the first non-digit (space, in your case).

Bash saves the captured number in ${BASH_REMATCH[1]} (since it was the first set of parenthesis); we send that number to printf in order to get it zero-padded to three digits, then append on the - for the rest of the rename.

The rename only happens if the match was successful (via the && chaining).

I strongly recommend using mv -i or -n when doing any sort of bulk rename like this, to avoid silently & irreversibly deleting files in case of a name conflict. — Gordon Davisson, Feb 01 '19 at 18:56
absolutely; additional options are to run it on some test files, and/or prepend the mv with an echo (though you'd have to look carefully to know where the filenames start and end) — Jeff Schaller, Feb 01 '19 at 18:59

score 1 · Answer 2 · answered Feb 01 '19 at 19:11

1

If you have the Perl-based rename command

$ rename -n 's/.*#(\d+).*/sprintf "#%03d - %s", $1, $&/e' *.TXT
rename(AAAA #12 SSSS.TXT, #012 - AAAA #12 SSSS.TXT)
rename(BBB #231 CDF.TXT, #231 - BBB #231 CDF.TXT)
rename(CDFSDAAAA #1 AAAASS.TXT, #001 - CDFSDAAAA #1 AAAASS.TXT)

(-n added for testing purposes).

answered Feb 01 '19 at 19:11

steeldriver

81,074

The other solution is more simple, but I will try this one too. – Ancucchi Feb 01 '19 at 19:14
https://unix.stackexchange.com/q/229230/117549 in case it helps distinguish what rename you have... – Jeff Schaller Feb 01 '19 at 19:17

Batch rename a lot of file using a number find in filename

digit digit digit space

2 Answers2