Regular Expression With 'match'

[kurianos3d]

Hi joojaa

Thaks thx alot for ur good mind to share and this help

reds from

o.S.k

[ewerybody]

Ahhhh!! cool!

Thanks a very lot for that enlighting insight! :]

that "finding-replacing"-thing was new to me. I aprechiate a lot that you show the patience to explain that to us!

[ewerybody]

hehe :] why so late?

I found a good reg-Expression How-To in Brian Evert's good ol How-To's: MEL How-To #94 Useful Regular Expressions for MEL
another man of genius!

[deane]

QUOTE(ewerybody @ 09/20/05, 05:33 AM) [snapback]215207[/snapback]
I found a line in one of Naughtys scripts :
CODE
match "[^|]*$" $x
Very useful if you just want to get the last part of a maya-path.
makes pPlane3 out of |pPlane1|pPlane2|pPlane3 for instance.

But how can I get everything but the last part??!

I don't think that there's any regular expression that can do it on its own, but you can get it using a combination of substring and match, as follows:

CODE
substring($s, 1, size($x) - size(match("|[^|]*$", $x)))

The code above uses match to find the last component of the path, including the separator, then uses the size of that to tell substring where to stop.

[joojaa]

I do

Yes there actualy is (the details of this was outlined in the post 2, please reread the therad its usefull for you, but since naughty realy wanted to work on OS paths isntead of maya paths that was also the answer)

Anyway:

CODE
match "^.*|" "|pPlane1|pPlane2|pPlane3";
// Result: |pPlane1|pPlane2| //

This is because the matcher is by default GREEDY! BTW this ones also on brians page!

or if you can use new perllike subst in maya 7 wich i prefeer bacause it can backsubst (im abit misusing backsubst here since the new 3 part subst has s bug in it as it cant subs empty strings)

CODE
substitute "|pPlane1|pPlane2|pPlane3 " "s/[^|]*$/$1/";
// Result: |pPlane1|pPlane2| //

PS if you want all but the rest - the last | this is acualy challenging! Try it its dooable

Think!

Think more!

Cheater heres one way :

CODE
substitute "|pPlane1|pPlane2|pPlane3 " "s/(.+)[|].+$/$1/";
// Result: |pPlane1|pPlane2 //

And without the possible preceeding |

CODE
substitute "|pPlane1|pPlane2|pPlane3 " "s/[|]*(.+)[|].+$/$1/";

Note in the perllike language | on itself means something totaly else its or. Note both of theese are available on the linked page!

[deane]

QUOTE(deane @ 06/30/06, 09:31 AM) [snapback]241255[/snapback]
I don't think that there's any regular expression that can do it on its own

QUOTE(Joojaa @ 07/01/06, 01:28 AM) [snapback]241281[/snapback]
Yes there actualy is (the details of this was outlined in the post 2, please reread the therad its usefull for you, but since naughty realy wanted to work on OS paths isntead of maya paths that was also the answer)

I did read the thread. No one had responded to the poster's original question which was whether there was a regular expression which could give him the path minus the final component. I'm well aware of the "greedy" nature of REs, but the expression that you provided generates an invalid Maya path because of the trailing '|'. Try executing 'ls |pPlane1|pPlane2|" in the Script Editor and you'll find that you get a syntax error. Yes, that can be subsequently stripped off in variety of different ways, but the essential point is that there is no regular expression which will do the job by itself. If you want to use something other than just a regular expression, then there are lots of different ways of doing it.

Since making my earlier post, it has occurred to me that perhaps the simplest solution is this:

CODE
substitute("|[^|]*$", "|group1|pPlane1|pPlaneShape1", "");
// Result: |group1|pPlane1 //

[ewerybody]

hey deane!!
Cool you hopped in :] and keep the thread alive. I mean its almost a year old
Joojaa did help me very very much!! And I posted the new post just to gather stuff about MEL-RegEx. I know there is nothing about the initial problem. But its about the topic.

So btw: I found a little app that I thought may help ... The Regex Coach
Don't know if you guys already know... but looks interesting.

And thanks for your suggestions! I really begin to understand. And joojaa you won't believe!!: Today my regular expressionc pocket reference actually helped me out for the first time in a question about greedyness!

[joojaa]

I have something similar to regex coach but as a pure JavaScript page on my girlfriends computer because she needed to learn regexps quickly so i whipped one up in afternoon.

Anyway if find the time to translate the lessons from Finnish to English i could post it. The JavaScript itself is very simple.

Deane yes the second to last example is the same as yours (its just a a full substitute witch i prefer, finding it easier to expand a back substitution regexp later so I always use the 3 part subst command over match)! But to point out the original question didn

[ewerybody]

QUOTE
...paragraph syntaxic analysis...
?!? No not me! I had russian and a electrical engineering -A-levels-certificate-thing-sort-of... but nothing like this in school.

About the highend-thing.. You mean its very highend-style to just post a question, get the answer and never again touch the thread?
I really enjoy discussing things like in here for example. But some questions are very short and demand a short answer.. several times I tried to start mel topics to just discuss something .. and sometimes it happens that no one cares..

[joojaa]

QUOTE(ewerybody @ 07/03/06, 07:06 PM) [snapback]241409[/snapback]
?!? No not me! I had russian and a electrical engineering -A-levels-certificate-thing-sort-of... but nothing like this in school.

Well yeah i suppose it has something to do with the fact that most of us can read and spell any word they have ever heard at age 8 (while i didnt read Finnish as my primary language the shooling practice did flow over, besides i had 2 primary languages anyway). So you can spend the next 10 years worth of lessons analyzing the synax of sentences, paragraphs and meanings. The primary consideration of the language is after all to be understood not to be fancy (in Finnish very pronoucedly so).

Having a spelling contest in Finnish mindboglingly stupid, compared to english for example. Hoever having a contest to identify the wordform is challenging.

QUOTE
About the highend-thing.. You mean its very highend-style to just post a question, get the answer and never again touch the thread?

Yes, nowdays thats so. Its actualy hard to get a thread running where the ideas bound freely. The best sort of discussion is realy hard to acheve on the net tough. Having a lively debatte seems hard since what happens is somebody takes it as a personal attack. and refuses to play anymore. But i suppose thets teh way it is.

Or maybe its because people dont ask the right questions....

[ewerybody]

ok finally resuming that thread... :]

I want to retrieve an edge number: This is what I have:
CODE
match "[[1-9]+]" "boden4.e[316]"
// Result: [316] //
Now how can I get rid of the [] but be certain to geht the number between the [] as well?!?! The fools idea was to loop it through another match "[1-9]+" "[316]". But I bet there is a way to do this in one go ;]

[joojaa]

substitute "boden4.e[316]" "s/.*\[([0-9]+)\]/$1/" ;
// Result: 316 //

just remember the regexp is a language of its own so tat line creates real code that may be much longer than doing this yourself so in this case it can actually faster to split with "[]" and get the last array item.

[ewerybody]

Ok. I'll give it all a try. But thanks a lot anyways!!! :]

to not spam it here I posted the thing that I was about here : Comparing Match, Substitute And Tokenize

[ewerybody]

Hey another regex-problem.

I need to test if a string has a *

I tryed kinds of escaping .. but everything seems to fail.

CODE
gmatch "pla*" "*";
// Result: 1 //
gmatch "pla" "*";
// Result: 1 //
gmatch "pla*" "[\*]";
// Result: 0 //

ahh I found a solutuion:
CODE
size(match("\","pla"));
// Result: 1 //
size(match("\*","pla"));
// Result: 0 //
But why the hell two \ ?!?

[joojaa]

The ""'s themselves are a escape sequence so \ -> \ to be shown inside "" off course your escaping seascapes, the bnext level is much weirder.

[ewerybody]

yay ok. I dug it :] thank you a lot!

(what I did was some kind of firefox like findWhenTyping: I have a scrollField that acts on keystroke its looking for objects. eg if you type "p" persp and top is automatically selected, But I wanted it to be able to select only persp on "p*" :] )

[underearth]

just a simple question.. if i want to match extension of maya file say ".mb"

and my scene name is delet_imb.mb

string $curSceneName =delet\_imb.mb;
string $newSceneName[];
string $numTokens = tokenize $curSceneName "[^.mb]$" $newSceneName;

can you explain whats wrong in this code..

where as

string $numTokens = tokenize $curSceneName "[^.]$" $newSceneName;

just did the job with "." for any file extension...

[joojaa]

wrong command tokenize dont accept regular expressions, it just splits at each supplied token. Th

So

CODE
string $curSceneName =delet\_imb.mb;
string $newSceneName[];
string $numTokens = tokenize $curSceneName "[^.mb]$" $newSceneName;

would return num tokens 1 and delet_imb because your lsit contains . m and b wich all define as tokenization split.

the second works for same reason, which is that you don't understand tokenize.

match "\..*$" "delet_imb.mb" would match any extension

[naughty]

sunny, you only use tokenize for spliting a string based on a single character, so if you give more than one character you're not giving a match sequence, you're giving it multiple single characters to spilt with...

besides, Maya offers many MEL commands and scripts for working specifically with filenames, so you could use a command like "baseExtension" (sorry, I don't have Maya to hand so I don't remember the actual command name but it's something like that... look in the docs for "baseName" and it should be referred to there!)

also, don't forget all matching functions are CaSe sEnsitive, so it's often useful to lower or UPPER case every string as you match or compare it... (tolower, toupper)

:nathaN

[ewerybody]

Unfortunately basename and basenameEx do not fit for this problem.
Actually I really never used them.

What does the job is the regExp from the very very first post of this thread.

QUOTE
// example
match "[^.]*$" "c:/qwdq/asgasdf.mb";
// Result: mb //
The credits go to Naughty nevertheless!! :]