'find' regular expression

[dragos2]

Dear all,

I have the find below:


find . -regextype posix-egrep -iname "*G[0-6]0*[1-4]*REV[1-9]*.pdf"

which returns files of certain revision number.

But I need the latest revision number in every directory and
this regular expression does not work as expected


*G[0-6]0*[1-4]*REV(9|8|7|6|5|4|3|2|1)*.pdf


The number trick should return the highest number without sort
but it does not work.

What do you think ?

[diesch]

You can't get only the highest number without comparing the numbers, and the regex can't do that as it only tries to match one filename at a time.

[Lars Noodén]

It might help to post a few examples of the file names to be sorted.

It's generally frowned upon to use "ls" in a script, but the following might work:



ls -v *.pdf | tail -n 1

[dragos2]

In every directory there should be files like this:
somecharacters-G101-REV2-sometext.pdf
somecharacters-G101-REV4-sometext.pdf

and I need the highest REV4 one from that directory.

Possible with find or grep ?

[Vaphell]

no. Regexes don't compare anything, they find matches. You need some kind of sort to define X>Y relation between returned results.

[dragos2]

no. Regexes don't compare anything, they find matches. You need some kind of sort to define X>Y relation between returned results.

I know, but I tried the REV(9|8|7|6|5|4|3|2|1).

But I think find operates on individual files and the above regex
can't work.

It seems that I have to use a real language for this.

[Vaphell]

but it didn't make any sense here
it was equivalent of [1-9] either way



some-junk: does it match REV[1-9]? NO
file-REV3: does it match REV[1-9]? YES
file-REV4: does it match REV[1-9]? YES
file-REV8: does it match REV[1-9]? YES

names that got YES will be included in the results. That's all.


It seems that I have to use a real language for this.

what do you mean by real language? What's wrong with sorting and picking the first/last element? sort lets you define arbitrary sorting order.

[Lars Noodén]

sort lets you define arbitrary sorting order.

sort -k can do a lot, but the manual page is not very clear at all.

[drmrgd]

What about something like this (assuming your name example above is the way you've really named your files):


$ ls
somechars-G101-REV1-sometext.pdf somechars-G101-REV3-sometext.pdf somechars-G101-REV5-sometext.pdf
somechars-G101-REV2-sometext.pdf somechars-G101-REV4-sometext.pdf somechars-G101-REV6-sometext.pdf
$ find . -type f -iname "*REV*" -exec ls {} \; | sort -t "-" -rV -k3
./somechars-G101-REV6-sometext.pdf
./somechars-G101-REV5-sometext.pdf
./somechars-G101-REV4-sometext.pdf
./somechars-G101-REV3-sometext.pdf
./somechars-G101-REV2-sometext.pdf
./somechars-G101-REV1-sometext.pdf

[dragos2]

What about something like this (assuming your name example above is the way you've really named your files):


$ ls
somechars-G101-REV1-sometext.pdf somechars-G101-REV3-sometext.pdf somechars-G101-REV5-sometext.pdf
somechars-G101-REV2-sometext.pdf somechars-G101-REV4-sometext.pdf somechars-G101-REV6-sometext.pdf
$ find . -type f -iname "*REV*" -exec ls {} \; | sort -t "-" -rV -k3
./somechars-G101-REV6-sometext.pdf
./somechars-G101-REV5-sometext.pdf
./somechars-G101-REV4-sometext.pdf
./somechars-G101-REV3-sometext.pdf
./somechars-G101-REV2-sometext.pdf
./somechars-G101-REV1-sometext.pdf

Interesting but does not work because there might be dashes before. But I know that its the last dash the one we need.

Can anything be improved upon that ?

File naming syntax:


*G[0-6]0*[1-4]*REV[1-9]*.pdf

[Lars Noodén]

Interesting but does not work because there might be dashes before. But I know that its the last dash the one we need.

Can anything be improved upon that ?

If the file names will take other forms, it will help to have as much of the full range of variation as possible. Are there other examples that you can provide?

[Vaphell]

yup, you should give an example as broken as possible. Doing easy stuff is easy, the devil is always in the details and they tend to gather around corner cases and irregularities.

[drmrgd]

I'm not quite sure I'm clear on what you mean by dash comes before. It seems you're saying you can't use a dash as delimiter in that sense. So, what if you just sort a grep statement:


find new/ -type f -iname "*REV*" -exec ls {} \; | grep -i "REV" | sort -rV
new/-this-is-a-file-too-string-REV9-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV8-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV7-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV6-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV5-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV4-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV3-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV2-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV1-I-should-do-too.pdf
new/somthingelse-aaa-rev7-something.pdf
new/somthingelse-aaa-rev6-something.pdf
new/somthingelse-aaa-rev5-something.pdf
new/somthingelse-aaa-rev4-something.pdf
new/somthingelse-aaa-rev3-something.pdf
new/somthingelse-aaa-rev2-something.pdf
new/somthingelse-aaa-rev1-something.pdf
new/idunno-abk-REV5-thistext.pdf
new/idunno-abk-REV4-thistext.pdf
new/idunno-abk-REV3-thistext.pdf
new/idunno-abk-REV2-thistext.pdf
new/idunno-abk-REV1-thistext.pdf
new/file1-xxx-rev3-somthing.pdf
new/file1-xxx-rev2-somthing.pdf
new/file1-xxx-rev1-somthing.pdf

[Lars Noodén]

$ find . -type f -iname "*REV*" -exec ls {} \; | sort -t "-" -rV -k3


If the REV # goes beyond 9, the sort needs to be modified.



| sort -t "-" -rV -k3.4


But that's just having fun with sort until there are more sample data.

[drmrgd]

I think the '-V' option takes care of that case:


find . -maxdepth 1 -type f -iname "*REV*" -exec ls {} \; | sort -t "-" -rV -k3./somechars-G101-REV15-sometext.pdf
./somechars-G101-REV14-sometext.pdf
./somechars-G101-REV13-sometext.pdf
./somechars-G101-REV12-sometext.pdf
./somechars-G101-REV11-sometext.pdf
./somechars-G101-REV10-sometext.pdf
./somechars-G101-REV9-sometext.pdf
./somechars-G101-REV8-sometext.pdf
./somechars-G101-REV7-sometext.pdf
./somechars-G101-REV6-sometext.pdf
./somechars-G101-REV5-sometext.pdf
./somechars-G101-REV4-sometext.pdf
./somechars-G101-REV3-sometext.pdf
./somechars-G101-REV2-sometext.pdf
./somechars-G101-REV1-sometext.pdf

Good to know you can do that with the -k option for future reference, though. Didn't know that was possible.

[dragos2]

I'm not quite sure I'm clear on what you mean by dash comes before. It seems you're saying you can't use a dash as delimiter in that sense. So, what if you just sort a grep statement:


find new/ -type f -iname "*REV*" -exec ls {} \; | grep -i "REV" | sort -rV
new/-this-is-a-file-too-string-REV9-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV8-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV7-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV6-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV5-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV4-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV3-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV2-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV1-I-should-do-too.pdf
new/somthingelse-aaa-rev7-something.pdf
new/somthingelse-aaa-rev6-something.pdf
new/somthingelse-aaa-rev5-something.pdf
new/somthingelse-aaa-rev4-something.pdf
new/somthingelse-aaa-rev3-something.pdf
new/somthingelse-aaa-rev2-something.pdf
new/somthingelse-aaa-rev1-something.pdf
new/idunno-abk-REV5-thistext.pdf
new/idunno-abk-REV4-thistext.pdf
new/idunno-abk-REV3-thistext.pdf
new/idunno-abk-REV2-thistext.pdf
new/idunno-abk-REV1-thistext.pdf
new/file1-xxx-rev3-somthing.pdf
new/file1-xxx-rev2-somthing.pdf
new/file1-xxx-rev1-somthing.pdf

I get a broken pipe on the ls part.

drmrgd try this: cp new new2 and try your code again.

I need only one file from one directory: the highest REV.

[Lars Noodén]

There might be a way to do this in awk, but after trying for a while I broke down and escalated to perl. This reads a list of files from stdin and expects the file names to have REV or rev plus some numbers somewhere in the file name.



#!/usr/bin/perl

use strict;

my %files = ();

# cycle through files provided on stdin
foreach my $file ( <> ) {
chomp( $file );
my @base = ();
my $revision = 0;

# cycle through file name components, separated by "-"
foreach my $part ( split( /-/, $file ) ) {

# grab revision number
if ( $part =~ /^rev(\d+)/i) {
$revision = $1;
next;
}
push( @base, $part);
}

my $base = join ('-', @base );

# keep the file if it is the most recent revision so far
if ( $files{$base}{'rev'} < $revision ) {
$files{$base}{'rev'} = $revision;
$files{$base}{'file'} = $file;
}
}

# print out the latest revisions
foreach my $base ( keys( %files ) ) {
print qq($files{$base}{'file'}\n);
}

exit (0);


It can be invoked like this.



find ./new -iname '*rev*' -print | ./latest_revision.pl

[drmrgd]

I get a broken pipe on the ls part.

drmrgd try this: cp new new2 and try your code again.

I need only one file from one directory: the highest REV.

Not sure what the broken pipe error is. If I copy new to new2 as you propose and run the command, I get the same results


cp -r new/ new2
$ find new2/ -type f -iname "*REV*" -exec ls {} \; | grep -i "REV" | sort -rV
new/-this-is-a-file-too-string-REV9-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV8-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV7-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV6-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV5-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV4-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV3-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV2-I-should-do-too.pdf
new/-this-is-a-file-too-string-REV1-I-should-do-too.pdf
new/somthingelse-aaa-rev7-something.pdf
new/somthingelse-aaa-rev6-something.pdf
new/somthingelse-aaa-rev5-something.pdf
new/somthingelse-aaa-rev4-something.pdf
new/somthingelse-aaa-rev3-something.pdf
new/somthingelse-aaa-rev2-something.pdf
new/somthingelse-aaa-rev1-something.pdf
new/idunno-abk-REV5-thistext.pdf
new/idunno-abk-REV4-thistext.pdf
new/idunno-abk-REV3-thistext.pdf
new/idunno-abk-REV2-thistext.pdf
new/idunno-abk-REV1-thistext.pdf
new/file1-xxx-rev3-somthing.pdf
new/file1-xxx-rev2-somthing.pdf
new/file1-xxx-rev1-somthing.pdf

or:


find . -type f -iname "*REV*" -exec ls {} \; | grep -i "REV" | sort -rV
./somechars-G101-REV15-sometext.pdf
./somechars-G101-REV14-sometext.pdf
./somechars-G101-REV13-sometext.pdf
./somechars-G101-REV12-sometext.pdf
./somechars-G101-REV11-sometext.pdf
./somechars-G101-REV10-sometext.pdf
./somechars-G101-REV9-sometext.pdf
./somechars-G101-REV8-sometext.pdf
./somechars-G101-REV7-sometext.pdf
./somechars-G101-REV6-sometext.pdf
./somechars-G101-REV5-sometext.pdf
./somechars-G101-REV4-sometext.pdf
./somechars-G101-REV3-sometext.pdf
./somechars-G101-REV2-sometext.pdf
./somechars-G101-REV1-sometext.pdf
./new/-this-is-a-file-too-string-REV9-I-should-do-too.pdf
./new/-this-is-a-file-too-string-REV8-I-should-do-too.pdf
./new/-this-is-a-file-too-string-REV7-I-should-do-too.pdf
./new/-this-is-a-file-too-string-REV6-I-should-do-too.pdf
./new/-this-is-a-file-too-string-REV5-I-should-do-too.pdf
./new/-this-is-a-file-too-string-REV4-I-should-do-too.pdf
./new/-this-is-a-file-too-string-REV3-I-should-do-too.pdf
./new/-this-is-a-file-too-string-REV2-I-should-do-too.pdf
./new/-this-is-a-file-too-string-REV1-I-should-do-too.pdf
./new/somthingelse-aaa-rev7-something.pdf
./new/somthingelse-aaa-rev6-something.pdf
./new/somthingelse-aaa-rev5-something.pdf
./new/somthingelse-aaa-rev4-something.pdf
./new/somthingelse-aaa-rev3-something.pdf
./new/somthingelse-aaa-rev2-something.pdf
./new/somthingelse-aaa-rev1-something.pdf
./new/idunno-abk-REV5-thistext.pdf
./new/idunno-abk-REV4-thistext.pdf
./new/idunno-abk-REV3-thistext.pdf
./new/idunno-abk-REV2-thistext.pdf
./new/idunno-abk-REV1-thistext.pdf
./new/file1-xxx-rev3-somthing.pdf
./new/file1-xxx-rev2-somthing.pdf
./new/file1-xxx-rev1-somthing.pdf
./new2/-this-is-a-file-too-string-REV9-I-should-do-too.pdf
./new2/-this-is-a-file-too-string-REV8-I-should-do-too.pdf
./new2/-this-is-a-file-too-string-REV7-I-should-do-too.pdf
./new2/-this-is-a-file-too-string-REV6-I-should-do-too.pdf
./new2/-this-is-a-file-too-string-REV5-I-should-do-too.pdf
./new2/-this-is-a-file-too-string-REV4-I-should-do-too.pdf
./new2/-this-is-a-file-too-string-REV3-I-should-do-too.pdf
./new2/-this-is-a-file-too-string-REV2-I-should-do-too.pdf
./new2/-this-is-a-file-too-string-REV1-I-should-do-too.pdf
./new2/somthingelse-aaa-rev7-something.pdf
./new2/somthingelse-aaa-rev6-something.pdf
./new2/somthingelse-aaa-rev5-something.pdf
./new2/somthingelse-aaa-rev4-something.pdf
./new2/somthingelse-aaa-rev3-something.pdf
./new2/somthingelse-aaa-rev2-something.pdf
./new2/somthingelse-aaa-rev1-something.pdf
./new2/idunno-abk-REV5-thistext.pdf
./new2/idunno-abk-REV4-thistext.pdf
./new2/idunno-abk-REV3-thistext.pdf
./new2/idunno-abk-REV2-thistext.pdf
./new2/idunno-abk-REV1-thistext.pdf
./new2/file1-xxx-rev3-somthing.pdf
./new2/file1-xxx-rev2-somthing.pdf
./new2/file1-xxx-rev1-somthing.pdf

That being said, I tried Lars' script and it's way nicer and gives much, much better output. Nice one! So, that's definitely the better way to go.


find ./new* -iname '*rev*' -print | ./latest_revision.pl
./new2/file1-xxx-rev3-somthing.pdf
./new2/-this-is-a-file-too-string-REV9-I-should-do-too.pdf
./new/file1-xxx-rev3-somthing.pdf
./new/-this-is-a-file-too-string-REV9-I-should-do-too.pdf
./new/idunno-abk-REV5-thistext.pdf
./new2/somthingelse-aaa-rev7-something.pdf
./new/somthingelse-aaa-rev7-something.pdf
./new2/idunno-abk-REV5-thistext.pdf