Regular Expressions

[lelandv]

Is someone able to sanity-check this regex for me?

It's designed to verify the format of a UK post code (in general formatting terms) - it does not do extended validation (i.e. if the first element has a letter suffix, it must have one or two letter prefixes and only one number etc.) but I think it should validate the general format.

^([a-z]|[A-Z]){1,2}(\d){1,2}([a-z]|[A-Z])?(\s)*(\d){1}([a-z]|[A-Z]){2}

The intended logic is:

Must begin with at least one letter but no more than 2, followed by one or two digits, followed by 0 or 1 letter;

may then have whitespace or none at all, followed by:

1 digit, and 2 letters.

Thanks.

Leland

[Fidgits]

the issue is, there can be 1 or 2 letters, before 1 or 2 numbers, followed by 2 letters..

but, it looks okay to me...

[lelandv]

Actually.. thinking about it.. could catch the extended format validation perhaps?

^([a-z]|[A-Z]){1,2}(\d){1}(\w)?(\s)*(\d){1}([a-z]|[A-Z]){2}

Feasible, no?

L.

the issue is, there can be 1 or 2 letters, before 1 or 2 numbers, followed by 2 letters..

hmm.. you mean in the first element?

I know you can have, for example:

WC1A

MK43

N18

W1

but I don't thinkyou can have:

W1AC

MK43B

W18A

and as far as I'm aware the last element is always a fixed format of one digit and two letters..

such as:

MK43 8ZZ

or

WC1A 3BQ

never seen anything different for the second element... or am I misguided?

Leland

[Steve ( West Mids )]

Does this help:

http://www.spot-light.virgin-lands.co.uk/postcode.php

[lelandv]

Thanks...

My final version is this:

^([a-zA-Z]){1,2}(\d){1}(\w)?(\s)*(\d){1}([a-zA-Z]){2}

which works in Perl, PHP, .NET (C#, J#, VB, C++), but I have a problem with ColdFusion...

Specifically the function:

REFind("^([a-zA-Z]){1,2}(\d){1}(\w)?(\s)*(\d){1}([a-zA-Z]){2}","WC1A 3LL")

Returns 0, meaning that it did not match at all. All the other languages match it correctly, and also pick up if an invalid format is used. ColdFusion, however, doesn't seem to want to play ball... though it accepts the syntax itself as "valid" without generating any errors...

Anyone have any ideas for coldfusion?

Thanks,

Leland

[eXOBeX]

Does Cold Fusion accept WC1A3LL (i.e. without the space)?

If so, it's a case of stripping all spaces, then parsing the postcode, then some sort of character shift along the lines of the following (written in Pascal in this case):-

postcode:=postcode+" "; {adds a space to the end, increasing length by 1}

postcode[length(postcode)]:=postcode[length(postcode)-1];

postcode[length(postcode)-1]:=postcode[length(postcode)-2];

postcode[length(postcode)-2]:=postcode[length(postcode)-3];

The last three lines nudge the last three characters along to recreate the gap.

[lelandv]

Does Cold Fusion accept WC1A3LL (i.e. without the space)?

If so, it's a case of stripping all spaces, then parsing the postcode, then some sort of character shift along the lines of the following (written in Pascal in this case):-

postcode:=postcode+" "; {adds a space to the end, increasing length by 1}

postcode[length(postcode)]:=postcode[length(postcode)-1];

postcode[length(postcode)-1]:=postcode[length(postcode)-2];

postcode[length(postcode)-2]:=postcode[length(postcode)-3];

The last three lines nudge the last three characters along to recreate the gap.

Nope.. CF doesn't accept any format with or without spaces... I suspect that there's something funny about CF's implementation of Regex :duh:

L.