# Split words into caracters **URL:** \ **Category:** NGINX\ **Created:** [April 2, 2013, 11:41am UTC](https://www.ruby-forum.com/t/split-words-into-caracters/227895 "2013-04-02T11:41:27Z")\ **Posts on this page:** 5\ **Page:** 1
**Author:** ![larry](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/l/47e85d/32.png) [@larry](https://www.ruby-forum.com/u/larry)\ **Post date:** [April 2, 2013, 11:41am UTC](https://www.ruby-forum.com/t/split-words-into-caracters/227895/1 "2013-04-02T11:41:27Z")
Hi ! I don’t know if it is the right way to go, and I hope to hear from you if it is not. My concern is about the selection of a file traversing a tree. I have a bunch of hashed directories (dozens of millions) on different servers. As for performance reasons, I would like to avoid having to many sub-directories in the same directory. I would rather have this scheme : 1234/5678/90AB/CDEF/1234/5678/90AB/CDEF/myfile It would help proxying to the right server and so on. Will the traverse be fast enought ? Every directory is approx 100kb heavy (really light). But other than openresty, i didn’t see any implementation inside nginx to split any word or cookie. C is faster than lua, so it would be a good bet. Well, I think. What is your word on it ? Thanks ! Larry Posted at Nginx Forum: > **[Split words into caracters](https://forum.nginx.org/read.php?2,238015,238015#msg-238015)** > > Hi ! I don't know if it is the right way to go, and I hope to hear from you if it is not. My concern is about the selection of a file traversing a tree. I have a bunch of hashed directories (dozens of millions) on different servers. As for... ---
**Author:** ![larry](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/l/47e85d/32.png) [@larry](https://www.ruby-forum.com/u/larry)\ **Post date:** [April 2, 2013, 11:47am UTC](https://www.ruby-forum.com/t/split-words-into-caracters/227895/2 "2013-04-02T11:47:18Z")
> But other than openresty, i didn’t see any implementation inside nginx to > split any word or cookie. You can possibly use rewrite [GitHub - nginxinc/nginx-wiki: ARCHIVED -- Source for the now archived NGINX Wiki section of https://www.nginx.com · GitHub](http://wiki.nginx.org/HttpRewriteModule) ( [Module ngx\_http\_rewrite\_module](http://nginx.org/en/docs/http/ngx_http_rewrite_module.html) ) There is also an example to rewrite /photos/123456 to something like /path/to/photos/12/1234/123456.png rr ---
**Author:** ![larry](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/l/47e85d/32.png) [@larry](https://www.ruby-forum.com/u/larry)\ **Post date:** [April 2, 2013, 5:39pm UTC](https://www.ruby-forum.com/t/split-words-into-caracters/227895/3 "2013-04-02T17:39:43Z")
> yeah but it works when you have multiple words. What do you mean by multiple words? > In my case, there is only one which is 1234567890…DEF (md5 → 32 chars) > m=1234567890…DEF > m[1] = 1 > m[2] = 2 > .. > m[32] = F > A mere rewrite is impossible here.. > It seems indeed, Why not? rewrite “/path/(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})(?[\w]{1})” /realpath/$r1/$r2/$r3/$r4/$r5/$r6/$r7/$r8/$r9/$r10/$r11/$r12/$r13/$r14/$r15/$r16/$r17/$r18/$r19/$r20/$r21/$r22/$r23/$r24/$r25/$r26/$r27/$r28/$r29/$r30/$r31/$r32; It looks somewhat ugly because of 32 variables (aso because you need to use named ones (since $10/$11 … wouldnt work)). There are probably much better ways to do this, but as a proof of concept it works: Request to [http://someserver/path/aBcdfghjklqwertyuiopas1234567890](http://someserver/path/aBcdfghjklqwertyuiopas1234567890) gets rewritten/translated to: 2013/04/02 18:25:37 [error] 2032#0: \*60 open() “/www/realpath/a/B/c/d/f/g/h/j/k/l/q/w/e/r/t/y/u/i/o/p/a/s/1/2/3/4/5/6/7/8/9/0” failed (2: No such file or directory), request: "GET /path/aBcdfghjklqwertyuiopas1234567890 (obviously error since such file doesnt exist for me) 🙂 rr ---
**Author:** ![larry](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/l/47e85d/32.png) [@larry](https://www.ruby-forum.com/u/larry)\ **Post date:** [April 2, 2013, 7:11pm UTC](https://www.ruby-forum.com/t/split-words-into-caracters/227895/4 "2013-04-02T19:11:27Z")
You are right Reinis, It first seemed tough to me but yeah, it works 🙂 Thanks, Larry Posted at Nginx Forum: > **[Re: Split words into caracters](https://forum.nginx.org/read.php?2,238015,238034#msg-238034)** > > Hi ! I don't know if it is the right way to go, and I hope to hear from you if it is not. My concern is about the selection of a file traversing a tree. I have a bunch of hashed directories (dozens of millions) on different servers. As for... ---
**Author:** ![larry](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/l/47e85d/32.png) [@larry](https://www.ruby-forum.com/u/larry)\ **Post date:** [April 2, 2013, 4:47pm UTC](https://www.ruby-forum.com/t/split-words-into-caracters/227895/5 "2013-04-02T16:47:42Z")
yeah but it works when you have multiple words. In my case, there is only one which is 1234567890…DEF (md5 → 32 chars) And I would need the possibility to do m=1234567890…DEF m[1] = 1 m[2] = 2 … m[32] = F A mere rewrite is impossible here… It seems indeed, Larry Posted at Nginx Forum: > **[Re: Split words into caracters](https://forum.nginx.org/read.php?2,238015,238029#msg-238029)** > > Hi ! I don't know if it is the right way to go, and I hope to hear from you if it is not. My concern is about the selection of a file traversing a tree. I have a bunch of hashed directories (dozens of millions) on different servers. As for...