c# UK postcode splitting

Question

I need a way to split a UK postcode from user entry. This means the postocode could be nicely formatted full code like so "AB1 1BA" or it could be anything you could imagine. I've seen some regex to check the format of the postcode but it's knowing where to split it if I'm given something like "AB111AD" etc.. This is to return the first part of the postcode, in the example above would be "AB11". Any thoughts? Thanks..

Answer 1

I've written something similar in the past. I think you can just split before the last digit. (e.g. remove all spaces, find the last digit and then insert a space before it):

static readonly char[] Digits = "0123456789".ToCharArray();

...

string noSpaces = original.Replace(" ", "");
int lastDigitIndex = noSpaces.LastIndexOfAny(Digits);
if (lastDigitIndex == -1)
{
    throw new ArgumentException("No digits!");
}
string normalized = noSpaces.Insert(lastDigitIndex, " ");
                           

The Wikipedia entry has a lot of detail including regular expressions for validation (after normalisation :)

Answer 2

I'm not sure how UK Post Codes work, so is the last part considered the last 3 characters with the first part being everything before?

If it is, something like this should work, assuming you've already handled appropriate validation: (Edited thanks to Jon Skeets commment)

string postCode = "AB111AD".Replace(" ", "");
string firstPart = postCode.Substring(0, postCode.Length - 3);

That will return the Post Code minus the last 3 characters.

Answer 3

UK-postcodes format explained:

Ref: http://www.mrs.org.uk/pdf/postcodeformat.pdf

POSTCODE FORMAT

A Postcode is made up of the following elements: PO1 3AX

• PO the area. There are 124 postcode areas in the UK
• 1 the district. There are approximately 20 Postcode districts in an area
• 3 the sector. There are approximately 3000 addresses in a sector.
• AX the Unit. There are approximately 15 addresses per unit.

The following list shows all valid Postcode formats. "A" indicates an alphabetic character and "N" indicates a numeric character.

FORMAT EXAMPLE:

AN NAA - M1 1AA
ANN NAA - M60 1NW
AAN NAA - CR2 6XH
AANN NAA - DN55 1PT
ANA NAA - W1A 1HQ
AANA NAA - EC1A 1BB

Please note the following:

• The letters Q, V and X are not used in the first position
• The letters I,J and Z are not used in the second position.
• The only letters to appear in the third position are A, B, C, D, E, F, G, H, J, K, S, T, U and W.
• The second half of the postcode is always consistent numeric, alpha, alpha format and the letters C, I, K, M, O and V are never used.

And it is safe to assume that the space will be the forth character from the end, ie., if a postcode is missing a space, SW109RL, you can blindly put a space at the 4th position from the end, SW10 9RL

Answer 4

I have worked with many UK insurance websites and we normally ask both the parts in different text boxes. How are you validating the address? In some sites we ask the post code together but we use QAS to validate the postcode and ask the user to select the address. QAS can validate even if the postcode is entered together.

Answer 5

Regular expressions may help to easily parse UK post code by using named groups for each part of the code. Regular expressions may be taken from here:

http://www.regxlib.com/REDetails.aspx?regexp_id=260

or here:

http://www.mgbrown.com/PermaLink66.aspx

string ukPostCode = "AB1 1BA";
// Add group names in the pattern like this {FIRST_GROUP}    
string UK_POST_PATTERN = @"^([A-PR-UWYZ0-9][A-HK-Y0-9][AEHMNPRTVXY0-9]?[ABEHMNPRVWXY0-9]? {1,2}[0-9][ABD-HJLN-UW-Z]{2}|GIR 0AA)$";

Regex ukPostRegex = new Regex(UK_POST_PATTERN, RegexOptions.Compiled);

Match match = ukPostRegex.Match(ukPostCode);
if (match.Success)
{
    Group group = match.Groups["FIRST_GROUP"];
   // etc
}

Answer 6

Assuming a valid input string, e.g. one that's passed the following regex:

"^[A-Za-z]{1,2}[0-9Rr][0-9A-Za-z]?\s*?[0-9][ABDEFGHJLNPQRSTUWXYZabdefghjlnpqrstuwxyz]{2}$"

Note: this allows optional space between the Outcode and Incode.

Then the following replacement regex will tidy it up - (splits it into two captures $1 and $2, then inserts the space):

postCode = Regex.Replace(postCode, "^(\S+?)\s*?(\d\w\w)$", "$1 $2")

Answer 7

I've written a method to handle this as best I can for my application.

Firstly I put a max length on my textbox @maxlength="8"

Then I call the following method in my controller:

    if (searchString != null)
    {
        searchString = HtmlHelpers.ParseStringToUkPostcode(searchString);
    }      

The method consists of:

public static string ParseStringToUkPostcode(string inputPostcode)
        {
            string outputPostcode = inputPostcode;

            outputPostcode = outputPostcode.Replace(" ", "").Trim();
            string outward = outputPostcode.Substring(0, outputPostcode.Length - 3);
            string inward = outputPostcode.Substring(Math.Max(0, outputPostcode.Length - 3));
            string postcode = string.Format("{0} {1}", outward, inward);
            outputPostcode = postcode;

            return outputPostcode;
        }

It's not pretty but it seems to be relatively robust for what I need and will handle the obvious variations of user provided post codes, in my area at least.

Disclaimer.

I don't need to look at postcodes outside my county as far as this app is concerned.

Answer 8

I wrote this regex based on @Jon Skeet's suggestion that last 3 are always digit-letter-letter

([a-zA-Z][a-zA-Z0-9]{1,3}) ?(\d[a-zA-Z]{2})

I will probably let through some bad ones due to this part [a-zA-Z0-9]{1,3} but hey its good enough for me. You can test it here.

Answer 9

According to guidelines the UK Postcode's second part is always 3 characters starting with a number thus my excessive exception throwing. You can add other checks like; "if greater than 8 chars" and "if contains any numbers". Up to you.

Also please feel free to change the extension name, I don't like it, its too long. Happy coding!

Postcode Guide Guide to UK Postcodes

Try the following code:

Add it as a string extension so the use becomes easy.

Use this link to do that Create String Extensions

public static string FirstPartOfPostcode(this string str)
    {
        string postCodeNoSpaces = str.Replace(" ", "");

        char lastDigit = postCodeNoSpaces[postCodeNoSpaces.Length - 3];

        if(!Char.IsDigit(lastDigit))
        {
            throw new ArgumentException("Invalid PostCode");
        }

        string firstPart = postCodeNoSpaces.Substring(0, postCodeNoSpaces.Length - 3);

        return firstPart;
    }

Use:

string postcode = "AB111AD";
string firstPart = postcode.FirstPartOfPostcode();

Answer 10

Brandon is correct, you can substring on the last 3 characters because this is the format defined by the standard BS 7666. So "AA1 11AA" isn't in fact a valid postcode.

Apparently the only breach of BS 7666 is British Forces Post Office postcodes, which begin BFPO and end with 4 numbers, e.g. "BFPO 1234".

The Wikipedia entry has more information, as well as regexs for validation. It also has a choice quote: "completely accurate validation is only possible by attempting to deliver mail to the address."

Answer 11

  Simply split the string in equally section or as your requirements it may be single/double/triple bit



string postCode = "AB111AD".Replace(" ", "");
    string firstPart = postCode.Substring(0, postCode.Length - 3);
    string lastPart = postCode.Substring(postCode.Length -3,postCode.Length);

Answer 12

My PHP function for this problem ....

function ParseStringToUkPostcode($inputPostcode)
{
    $outputPostcode = trim(str_replace(' ', '', $inputPostcode));
    $firstPart = substr($outputPostcode, 0,strlen($outputPostcode)-3);
    $lastPart = substr($outputPostcode,strlen($outputPostcode)-3);
    $outputPostcode = $firstPart.' '.$lastPart;
    return $outputPostcode;
}