Extract certain part of filename in Bash

Question

I have many files in a folder:

yyyymmdd_hhmmss.mp4
yyyymmdd_hhmmss_suffix1.mp4
yyyymmdd_hhmmss_suffix1_suffix2.mp4

The following filename formats are also possible (rarely):

yyyymmdd_hhmmss_$$$.mp4
yyyymmdd_hhmmss_$$$_suffix1.mp4
yyyymmdd_hhmmss_$$$_suffix1_suffix2.mp4
yyyymmdd_hhmmss_$$.mp4
yyyymmdd_hhmmss_$$_suffix1.mp4
yyyymmdd_hhmmss_$$_suffix1_suffix2.mp4
yyyymmdd_hhmmss_$.mp4
yyyymmdd_hhmmss_$_suffix1.mp4
yyyymmdd_hhmmss_$_suffix1_suffix2.mp4

where $ is a number 0-9

I am trying to catch "yyyymmdd_hhmmss" and use it as an argument. This is what I do when only one suffix presented:

for file in "$@"; do 
  file_nosuffix="${file%*_suffix1.mp4}.mp4"
  echo "$file and $file_nosuffix"
done

But I get lost when all sorts of the filename formats mentioned above are presented. Ideally I would like to stick to the current pattern:

for file in "$@"; do 
   #catch "yyyymmdd_hhmmss"
   #do something on files yyyymmdd_hhmmss.mp4
   #do something else on files yyyymmdd_hhmmss_suffix1.mp4
   #etc.
done

Is that possible?

why don't you just save the first 15 characters into a variable? — Kathara
– Kathara, Commented Apr 7, 2017 at 15:07
Also, as an aside, it's helpful to provide test data that actually matches your stated format, rather than something that's descriptive to humans but can't be used for actual testing. yyyymmdd_hhmmss doesn't match something that expects all those characters to be digits, for example, which is presumably your actual format. — Charles Duffy
– Charles Duffy, Commented Apr 7, 2017 at 15:12
As a reference to get a substring: thegeekstuff.com/2010/07/bash-string-manipulation — Kathara
– Kathara, Commented Apr 7, 2017 at 15:12
Actually, I'd suggest BashFAQ #100 (part of the Wooledge wiki, maintained by the denizens of the irc.freenode.org #bash channel) for a discussion of string manipulation in bash over some random website. wiki.bash-hackers.org/syntax/pe is also reputable (and actively maintained). — Charles Duffy
– Charles Duffy, Commented Apr 7, 2017 at 15:13

Charles Duffy · Accepted Answer · 2017-04-07 19:56:44Z

4

Bash has built-in regex support, if you want to confirm the format:

regex='^[[:digit:]]{8}_[[:digit:]]{6}' # POSIX ERE; can't use PCRE extensions here

for file; do
  if [[ $file =~ $regex ]]; then
    echo "${BASH_REMATCH[0]} is the substring for $file" >&2
  else
    echo "$file does match the required format" >&2
  fi
done

One can also trivially take a prefix;

for file; do
  prefix=${file:0:15}
  echo "Prefix for $file is $prefix"
done

...or, to delete the last two underscores and everything after them:

prefix=${file%_*_*}

See:

BashFAQ #100 ("How can I do string manipulations in bash?")
Parameter Expansion

edited Apr 7, 2017 at 19:56

answered Apr 7, 2017 at 15:08

Charles Duffy

299k43 gold badges442 silver badges498 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

AndriusWild Over a year ago

Thank you all for your help. I think I am going to go ahead and take a prefix. However I wonder if there is a way to write code that will work for anyname_suffix1_suffix2.mp4

Charles Duffy Over a year ago

What logic do you actually want? Just kill _*_* from the end? prefix=${file%_*_*}, then.

AndriusWild Over a year ago

prefix=${file%__} is brilliant, that is exactly what I want!! Thank you @Charles Duffy

Charles Duffy Over a year ago

Amended, then. (You can mark your question answered via the checkbox next to the answer).

Collectives™ on Stack Overflow

Extract certain part of filename in Bash

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related