confused about regex, need some help

[b4t3m4n]

Hey guys, the thing I always find myself needing from a REGEX is pulling data from a form like

Artist: Some Artist
Album: Some Album
Song: Some song

I want a regex that will run through a file containing entries like this and spit out all the Artists in the file for example. Every time I try to make my own, it works, but it also prints out the Artist: tag along with the actual artist's name which is no use too me. I just want the actual name of the artist, not the "Artist:" tag along with it. I am using grep for this, or maybe there is a better way, I dunno. This seems like it should be really simple, I am just really bad with regex.

Thanks for any help

[HalPomeranz]

The easiest thing for you to do is to just pipe the output of your grep command through sed:



grep ... | sed -r 's/^[^:]+:\s+//'


On each line of output, the above sed expression will remove everything up to and including the first colon plus any whitespace after the colon. So a line like "Artist: some string" should get converted to just "some string".

[ghostdog74]

since you data file is structured, meaning there are at least recognisable delimiters and standard format, using awk or equivalent that works well with fields is more appropriate tool then grep or sed



awk -F":" '/Artist/{print $2}' file

[HalPomeranz]

The problem with your awk expression is that it's going to leave leading whitespace at the beginning of each line, i.e., "Artist: some string" is going to be converted to "<whitespace>some string". I was going to suggest using awk myself, but couldn't come up with a reasonable awk expression that would remove the header+whitespace from each line (I could do it with a Perl one-liner, but it would be messy to type). sed was easier.

[ghostdog74]

that's just trivial


awk -F": " '/Artist/{print $2}' file

[WW]

Or even...


awk -F":[ \t]*" '/Artist/{print $2}' file


However, using the field separator this way will cause a problem if there is a colon in the data string.

[b4t3m4n]

thanks for all the posts guys, I will try some of this and see what I can get to work for me

[HalPomeranz]

Hmmm, I realized there's another issue with the awk solution. What if there are colons later in the line? For example a line like "Album: Journey: Greatest Hits". The awk is just going to spit out $2 or "Journey" in this case, which is probably not what you want.

[ghostdog74]

Hmmm, I realized there's another issue with the awk solution. What if there are colons later in the line? For example a line like "Album: Journey: Greatest Hits". The awk is just going to spit out $2 or "Journey" in this case, which is probably not what you want.

you are getting way ahead. The awk solution is based on what sample data OP has given. It seemed structured to have only 1 colon. If its what you described, then it also trivial to modify to suit OP's need


awk -F"[ \t]*:[ \t]*" '/Artist/{$1="";sub(/^ +/,"");print}' file

[WW]

Instead of using the colon as the field separator in the awk command, you could process the line with the sub function before printing it:


awk "/^Artist: */ {sub(\"^Artist: *\",\"\"); print}" file

More generally,


export FIELD=Song
awk "/^${FIELD}: */ {sub(\"^${FIELD}: *\",\"\"); print}" file

(I assume that only spaces occur between the first colon and the start of the name. Change : * to :[ \t]* if tabs can also occur.)

For example:


$ cat file
Artist: Some Artist
Album: Some Album
Song: Some song
Artist: A. Artist
Album: Awkward Arguments
Song: Portrait of the Artist: My Story
$ export FIELD=Artist
$ awk "/^${FIELD}: */ {sub(\"^${FIELD}: *\",\"\"); print}" file
Some Artist
A. Artist
$ export FIELD=Song
$ awk "/^${FIELD}: */ {sub(\"^${FIELD}: *\",\"\"); print}" file
Some song
Portrait of the Artist: My Story
$



EDIT: Or you could use sub like ghostdog74 did. :)

[WW]

If awk makes you feel awkward, you could use perl:


perl -n -e 'if (/^Artist: *(.*)/) {print "$1\n"}' file

For example,


$ cat file
Artist: Some Artist
Album: Some Album
Song: Some song
Artist: A. Artist
Album: Awkward Arguments
Song: Portrait of the Artist: My Story
$ perl -n -e 'if (/^Artist: *(.*)/) {print "$1\n"}' file
Some Artist
A. Artist
$ perl -n -e 'if (/^Song: *(.*)/) {print "$1\n"}' file
Some song
Portrait of the Artist: My Story
$