LinuxQuestions.org - bash reg expression to split many file names on _3charnumbers

- Linux - General (https://www.linuxquestions.org/questions/linux-general-1/)

- - bash reg expression to split many file names on _3charnumbers_ (https://www.linuxquestions.org/questions/linux-general-1/bash-reg-expression-to-split-many-file-names-on-_3charnumbers_-4175635540/)

bash reg expression to split many file names on _3charnumbers_

friends,

I think what I need to do is split strings with a regular expression

where : _ and any 3 numbers underscore is my deliminator

I need to split a string ( a dir full of file names , thousands of mov files )

they all have the various forms of this
jug_WNYW_0000013_001_CommercialBlock7.mov
jug_WNYW_0000013_002_CommercialBlock7.mov
all the way to
jug_WNYW_0000013_099_CommercialBlock7.mov
and here are other examples
jug_WNYW_0000013_041_WeatherTwo.mov
jug_WLWT_0000099_022_reds.mov
jug_WLWT_0001234_001_reds.mov
jug_WCPO_0008994_007_shewalkedintothedoor.mov
b99456666_007_shetakesawall.mov
c9902345_003_chocolatepieisgood.mov

what I really wish i could do get get what ever is LEFT of the underscore 3 wide number underscore

goodStuff_NNN_
i wish for the good stuff

I tried a reg expression with
tapeName="$(cut -d'_[0-9]{3}_' -f1 <<<"$jug_1978-03-31_wlwt_AIR_005_SubheadsTeasers")"

i Thought the _[0-9]{3} would be
underscore and any 3 numbers and underscore

I just want jug_1978-03-31_wlwt_AIR

I have had limited success with

$ echo "jug_1978-03-31_WDTN_AIR_005_SubheadsTeasers" | awk -F'_' '{print $1"_"$2"_"$3}'
returns
jug_1978-03-31_WDTN

and you see that will never work with a string like
c9902345_003_chocolatepieisgood.mov

so I am back to needing the LEFT of _NNN_
that is an increment _001_ to _999_

I have done some sed & awk .... but never reg expressions

also this is on a mac in bash....

it *seems the mac (os 10.11) also does not take sed from standard in .... isn't that nice

any assistance is appreciated - thank you

best regards

and THANK YOU

zimbot

Do you need to count the files or just trim what is to the right of the triple digit number?

If you are limited to bash on OS X then parameter expansion might be relevant.

http://wiki.bash-hackers.org/syntax/pe

Otherwise it is easy with the perl-based version of the rename utility.

if you're going from left to right

Code:

userx@slacurr.ent.org:~

$ fun=jug_WNYW_0000013_001_CommercialBlock7.mov



#furthest to the right 

$ sofun=${fun##*[0-9]_}



$ echo $sofun

CommercialBlock7.mov

Code:

userx@slacurr.ent.org:~

$ Name=jug_WLWT_0000099_022_reds.mov ;

 

$ leadingFname=${Name%%_[0-9]*} ;

 

$ endingFname=${Name##*[0-9]_} ;

 

$ newFileName="$leadingFname"_"$endingFname" ;

 

$ echo $newFileName

jug_WLWT_reds.mov

look under Substring Removal

Quote:

Originally Posted by zimbot (Post 5886976)

it *seems the mac (os 10.11) also does not take sed from standard in .... isn't that nice

I find that very hard to believe. But I wouldn't be caught dead using a mac, so I can't check.

Could be done with parameter substitution, but I always mess up complex globs. Does anyone know if/how you can include a (specific) repetition count in a glob in a substring removal like that ?.
I also agree the perl rename is best option as you can use regex to get the job done. But you'd better get it correct (for all cases).

Perl

Code:

$var1="jug_1978-03-31_wlwt_AIR_005_SubheadsTeasers";

$var2=(split(/_[0-9]{3}_/,$var1))[0];

print "var2 = $var2\n";



# o/p

var2 = jug_1978-03-31_wlwt_AIR

String mangling is one of Perl's strengths... :)

thanks everybody

I will be giving some of this a try

I very much appreciate the perl suggestion .
due to convience I hope for a full bash solution

thanks again