g1smd
msg:4410438 | 9:40 pm on Jan 24, 2012 (gmt 0) |
[A-Z] matches upper case
[a-z] matches lower case
[0-9] matches a digit
{2} matches two of the previous character grouping
{3} matches three
/ matches a literal slash
^ and $ are the anchoring.
|
Matthew1980
msg:4410444 | 9:59 pm on Jan 24, 2012 (gmt 0) |
Hi there,
So something like ^[A-Z]/[A-Z][0-9]{03}/{02}$ would be about right? I'll try that now.
Thanks for the pointer.
Cheers,
MRb
|
g1smd
msg:4410446 | 10:06 pm on Jan 24, 2012 (gmt 0) |
No.
The counts are used directly after the type of thing you want to match i.e. the character groups, otherwise the character groups each match only a single character.
|
Matthew1980
msg:4410448 | 10:11 pm on Jan 24, 2012 (gmt 0) |
My appologies for seeming dumb at this, but I've never really had to use it until I was giving this project.
[0-9{3}]
I've split it into smaller chunks and I've only to get this part to work, I really appreciate this help (I'll assume as you got my tweet just..)
Cheers once again,
MRb
|
g1smd
msg:4410449 | 10:35 pm on Jan 24, 2012 (gmt 0) |
[0-9{3}] matches a single instance of 0 1 2 3 4 5 6 7 8 9 { or }
[...] is a character group - it lists the characters that are allowed.
{2} says how many occurences of those characters it should match.
/[0-9]{2} matches a slash followed by two digits.
|
Matthew1980
msg:4410458 | 10:57 pm on Jan 24, 2012 (gmt 0) |
Hi there,
Thanks for your help, I ended up with this: [A-Z]/[A-Z]{2}[0-9]{3}/[0-9]{2}
Works perfectly.
Cheers,
MRb
|
g1smd
msg:4410461 | 11:12 pm on Jan 24, 2012 (gmt 0) |
If your pattern is start-anchored then it will match only a single character before the first slash. Your example showed two characters.
If there is no start anchoring it will match any number of characters before the first slash, but only if the last one before the slash is an upper case letter.
If there is no end anchoring it will match any number of characters after the final slash and two digits have been matched.
|
Matthew1980
msg:4410562 | 6:02 am on Jan 25, 2012 (gmt 0) |
Hi g1smd,
Well thanks for pointing that out, I hadn't noticed that last night.
This is what I have ended up with, works perfectly :) ^[A-Z]{2}/[A-Z]{2}[0-9]{3}/[0-9]{2}$, I guess that this can be further optimised using some \d and \w tags in there, but until I read more up on this, this will do just fine.
Cheers for all your help.
MRb
|
g1smd
msg:4410581 | 7:45 am on Jan 25, 2012 (gmt 0) |
Yah! You got there in the end, but the important point is that you probably now understand your code.
One thing about testing. When you test some code, don't just try values which you expect to work, also try a whole bunch of things that should not work and make sure that is actually the case. Try various combinations of invalid characters (even punctuation) and both too many and too few characters.
Finally, although you used [A-Z] in the code, if the actual range of allowed letters is small you could list those instead, e.g. ^[DNST][RTX]/ etc.
|
Matthew1980
msg:4410802 | 7:40 pm on Jan 25, 2012 (gmt 0) |
Hello g1smd,
Cheers! And yes you're right, my comprehension of regex (even though a simple expression) is now better for it! Yep I sat there until the wee hours testing every permutation of the string as I could think of and the pattern works great. I shall just stream line it by doing something with the \d instead of [0-9].
In this particular context, designed through vb.net I have complete control over what chars/ints are allowed in the text area via KeyPressed event.
Excellent, and again, thanks. I did tweet you with a thanks and WebmasterWorld tag ;)
Regards,
MRb
|
g1smd
msg:4410811 | 7:54 pm on Jan 25, 2012 (gmt 0) |
Cheers! IMO, [0-9] is more portable, and not worth changing to \d here.
Didn't see anything on Twitter.
|
|
