Re: one-line split file sequence checker

[jwkrahn]

"tighter" sure, change  grep(s/^.+?([0-9]+)$/$1/, to  map(/.([0-9]+)$/, and change  `ls *[0-9]` to  <*[0-9]> and change  grep(s/$/\n/, to  map("$_\n",.

[kyle]

You can make map(/.([0-9]+)$/, into map(/.(\d+)$/, too.

[wdef2]

perl -e 'for $c (sort {$a<=>$b} map(/.(\d+)$/, <*[0-9]>)){ ++$n == $c || print map("$_\n", $n...$c-1) ; $n=$c }'

Ah, that's getting closer to the line noise effect I was looking for, thanks ;=) Time to read up on that map command and why <> is globbing filenames in the current dir. I'm almost pleased to see there's no offered improvement to the basic idea. There's no way to get rid of that 'for' iterator, or the 'print'? -lane ? Any shorthand for sort {$a<=>$b}? Just wondering.

[kyle]

If you're just golfing, you can throw away a few characters:

perl -e 'for$c(sort{$a<=>$b}map/(\d+)$/,<*[0-9]>){++$n==$c||print map"
+$_\n",$n..$c-1;$n=$c}'

If you want to get rid of the sort, try an array:

perl -le '@f=map/(\d+)$/,<*[0-9]>;@x[@f]=@f;for(1..$#x){$x[$_]||print}
+'

Note that if you have a file named '0', it will basically ignore it, but that was the original behavior too. This may eat more memory, but I doubt that will be an issue until you're working with millions of files.