Re^3: Efficient walk/iterate along a string

[Anonymous Monk]

Grandfather,

Hope I've got the techno details and the "protocol" right here.

I replied to your repy on my main thread as below. Appreciate if you could elucidate (love that word!) the regexp as per my words below. Various explorations with variants on ?<= give me the protein string split on the KR characters, with or without null strings, strings of just K or R, strings where consective K/R were condensed to just one, or none. I'm sure I can work it out with about 5 hours of exploration, but if you could just put into words what the ?<= is doibg I'd ber grateful.

Thanks, GE.

Cheers, GE

-------------

Monks
I tried the split/tr/map approach suggested by Grandfather and got ~ 60% reduction in time, great ! (and yes you got my app requirement exactly right)

I think should withdraw my words that there's nothing to be gained from the special functions when you're just marching down strings and doing a seemingly arbitrary logic function (and also my words about inadequate context). Splitting at simple marker points and counting a simple set of chars within the splits, combined with Perl's superb list handling, is not arbitrary spaghetti logic. Lesson learned.

One query: the ?<= construct was new to me, I've played with it and read up on it (perlre/Extended Patterns/Look-Around Assertions) and I'm 80% there but if anyone could explain exactly how this RE is ensuring I get exactly what I need I'd be grateful.

Will continue to check some other posts, thanks all

GJE

20101206 Janitored by Corion: Added formatting, code tags, as per Writeup Formatting Tips