Converting uppercased 18-digit Id to valid Id

Question

I am currently facing an issue where an external system passes back an 18-digit Id that our platform has previously sent to them. However in the process the capitalization of the Id has been removed - read everything has become uppercase. I would imagine somebody has faced this problem before, but apparantly this is not the case.

So in short; is somebody aware of any existing (apex-) function that converts uppercased 18-digit Ids to a valid Salesforce Id?

Please beware; while a lot of answers exist for converting 15 to 18 digits, which is trivial, this is a different problem.

Answer 1

I ported Daniel Ballinger's answer here to Apex:

static String CHARS = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ012345';

static List<Boolean> getBitPattern(String c)
{
    Integer index = CHARS.indexOf(c.toUpperCase());
    List<Boolean> result = new List<Boolean>();
    for (Integer bitNumber = 0; bitNumber < 5; bitNumber++)
        result.add((index & (1 << bitNumber)) != 0);
    return result;
}
static String repairCasing(String input)
{
    if (String.isBlank(input) || input.length() != 18) return input;

    List<Boolean> toUpper = new List<Boolean>();
    toUpper.addAll(getBitPattern(String.valueOf(input.substring(15, 16))));
    toUpper.addAll(getBitPattern(String.valueOf(input.substring(16, 17))));
    toUpper.addAll(getBitPattern(String.valueOf(input.substring(17, 18))));

    String output = '';
    for (Integer i = 0; i < 15; i++)
    {
        String c = String.valueOf(input.substring(i, i+1));
        output += toUpper[i] ? c.toUpperCase() : c.toLowerCase();
    }
    output += input.substring(15, 18).toUpperCase();
    return output;
}

I tested it and it worked:

Id value1 = Id.valueOf('00129000007Kbn7AAC');
Id value2 = Id.valueOf('00129000007KBN7AAC');

system.assertEquals(value1, repairCasing(value2));

Answer 2

Parallel to Adrian's response I also cooked up my own version for reference;

public static Id correctCapitalizedId(String input) {
    String keyString = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ012345', result = '';
    Integer i = 0, s = 15;

    if (input.length() != 18) {
        return null;
    }

    for (String seq : input.split('(?<=\\G.{5})')) {
        Integer chk = keyString.indexOf(input.mid(s++,1));
        i = 0;
        for (String chr : seq.split('')) {
            result += seq.length() != 5  ? '' : (chk & (i == 0 ? ++i : (i*=2))) == 0 ? chr.toLowerCase() : chr;
        }
    }
    return Id.valueOf(result);
}

Answer 3

Given that the 18 character form of an ID is designed to be a unique identifier irrespective of the casing, I'm not sure why you would ever need to convert to a 15 digit ID.

Why not stay with the 18 character values? If you did for example need to use them as string map keys just normalise their case by using toLowerCase() on them. They remain valid ID values and can be converted to the type Id using Id.valueOf.

Answer 4

In case anyone else is using it, I took Adrian Larson's apex code and converted it to Python.

def getBitPatterns(c):
    CHARS = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ012345'
    index = CHARS.find(c)
    result = []
    for bitNumber in range(0,5):
        result.append((index & (1 << bitNumber)) != 0)
    return result

def repairCasing(inp):
    if len(inp) < 18:
        return 'Error'
    toUpper = []
    toUpper.append(getBitPatterns(inp[15:16]))
    toUpper.append(getBitPatterns(inp[16:17]))
    toUpper.append(getBitPatterns(inp[17:18]))
    toUpper = [item for sublist in toUpper for item in sublist]

    output = ''

    for i in range(0,15):
        c = inp[i:i+1]
        if toUpper[i]:
            output += c.upper()
        else:
            output += c.lower()

    output += inp[15:18].upper()

    return output