How to extract string from filename

Jude DaShiell jdashiel at
Wed Jul 29 15:57:27 UTC 2015

Something like pid~=[bp][a-zA-Z0-9] in bash could extract a pid from a 
filename.  The ~= operator tells bash to use regular expressions for 
matching.  All of this would probably have to be part of a for loop to 
get that done.

On Wed, 29 Jul 2015, Rob Harris wrote:

> Date: Wed, 29 Jul 2015 10:39:19
> From: Rob Harris <robh at>
> Reply-To: Linux for blind general discussion <blinux-list at>
> To: Linux for blind general discussion <blinux-list at>
> Subject: Re: How to extract string from filename
> Linux has or can have apache server running to give access to PHP, and I can
> see that working better using the explode(string, separator-char) command
> using the _ as separator,
> zz=explode(string, char);
> this gives an array of sub-string without the _'s,  and you need the second
> to last.  There's another function to count elements in an array,  so x-1
> where x is the number of elements.  Might be x-2 as first element is 0.
> I oculd get more specific if I dipped into all my old, but still used, PHP
> scripts.
> implode(array, separator) like zz2=implode(array, char), can put it back as
> a long string separated by the char.
> I love the relative simplicity of PHP for all this.
> It can suck in an entire file into a string then save the modified string to
> a new, or back to the old, file.
> You'd use .mp3 as separator in this first instance.
> Hth, BobH.
> ----- Original Message -----
> From: "Tony Baechler" <tony at>
> To: "Linux for blind general discussion" <blinux-list at>
> Sent: Wednesday, July 29, 2015 12:13 PM
> Subject: How to extract string from filename
> Hi all,
> The recent discussion on shell scripts got me thinking.  A couple of posters
> invited people to post problems they're having with scripts to the list, so
> here goes.
> I have not actually written a script for this because I'm not sure how to go
> about it.  I would normally use cut, but I need to cut from right to left.
> The cut help doesn't indicate a way to do this.  You can only cut from the
> beginning of the line or a range of bytes.  The problem is each line
> (filenames, to be exact) are of different lengths, so it's impossible to
> know what range of bytes I need.
> What I'm trying to do is extract the BBC PID from the downloaded files.
> It's a lower case alphanumeric string which starts with a letter and is
> eight characters.  In my case, the first letter is always "b" or "p," so if
> I could use something like grep to just extract the first lower case letter
> followed by a number up to the next underscore, that would be good.  I don't
> think grep will just print a matching phrase, only the matching line.  Here
> are some example filenames:
> 5_live_Science_-_Coding_and_Computers_b062dj5j_default.mp3
> Witness_-_The_Sinking_of_the_USS_Indianapolis_p02wdykn_default.mp3
> Discovery_-_A_Scientific_View_of_Agriculture_p0053gbd_default.mp3
> Click_-_05_10_2010_p00b18gp_default.mp3
> As you can see, they all follow a similar format.  If I could go from right
> to left, I would simply cut "_default.mp3" and extract the preceeding 8
> bytes, but I can't figure out how.  What I'm trying to do is first extract
> the PIDs, hopefully preserving the filenames in the process.  Once they are
> extracted (or printed to stdout), I want to use wget to download the BBC
> programme page.  If you go to, you'll get a
> web page displaying the broadcast date, description and notes.  I would like
> to download those pages.
> Any help with this would be greatly appreciated.  Thanks in advance.
> --------------------
> Tony Baechler, Baechler Access Technology Services
> Putting accessibility at the forefront of technology
> mailto:bats at
> Phone: 1-619-746-8310   Fax: 1-619-449-9898
> _______________________________________________
> Blinux-list mailing list
> Blinux-list at
> _______________________________________________
> Blinux-list mailing list
> Blinux-list at


More information about the Blinux-list mailing list