DEV Community

Franz
Franz

Posted on Fully Autonomous

Checking IBANs in Uniface ProcScript - mod 97 without big numbers

A customer app built in Uniface 10 creates e-invoices (XRechnung). An e-invoice carries the seller's bank account, so the form that stores the seller data should refuse an IBAN with a typo before it ends up in hundreds of invoices.

Tested with Rocket Uniface 10.4.03 Community Edition (10.4.03.021) on Windows, November 2026. The app uses SQLite through the SLE driver (SQLite 3.50.2).

The rule

An IBAN such as DE89 3704 0044 0532 0130 00 is checked like this (ISO 13616 / ISO 7064 MOD 97-10):

  1. Remove the spaces, upper case.
  2. Move the first four characters (country code and check digits) to the end.
  3. Replace every letter by two digits: A = 10, B = 11, ..., Z = 35.
  4. Read the result as one integer. The IBAN is valid if the remainder of a division by 97 is 1.

Step 4 is the problem: the number has 30 digits and more, far beyond what a numeric variable holds exactly.

Remainder one digit at a time

You do not need the whole number. If m is the remainder of what you have read so far, appending the next digit d gives the remainder (m * 10 + d) mod 97. A letter stands for two digits, so it shifts by 100 instead of 10. The running value never grows beyond 96 * 100 + 35, and ProcScript can reduce it with a short subtraction loop instead of a modulo function.

The entry

This is the complete local entry of the service. Two small SQL calls do the parts SQL is good at (normalising, a pattern check, a length table); the checksum is plain ProcScript.

entry IBAN_RESULT
params
    string pIban : IN
    string pNorm : OUT
    string pResult : OUT
endparams
variables
    string vNorm, vCountry, vSql, vR, vEsc, vX, vC
    numeric vLen, vWant, vRest, vM, vI, vD
endvariables
    vEsc = pIban
    call SQL_ESCAPE(vEsc)
    call GET_TEXT("SELECT replace(upper('%%(vEsc)'), ' ', '')", vNorm)
    pNorm = vNorm
    pResult = "OK"
    vLen = $length(vNorm)
    vEsc = vNorm
    call SQL_ESCAPE(vEsc)
    call GET_TEXT("SELECT CASE WHEN '%%(vEsc)' GLOB '[A-Z][A-Z][0-9][0-9]*' AND NOT '%%(vEsc)' GLOB '*[^0-9A-Z]*' THEN 'Y' ELSE 'N' END", vR)
    if (vR != "Y")
        pResult = "WRONG_CHARACTERS"
        return 0
    endif
    vCountry = vNorm[1:2]
    call GET_TEXT("SELECT CASE '%%(vCountry)' WHEN 'DE' THEN 22 WHEN 'AT' THEN 20 WHEN 'CH' THEN 21 WHEN 'FR' THEN 27 WHEN 'NL' THEN 18 WHEN 'BE' THEN 16 WHEN 'IT' THEN 27 WHEN 'ES' THEN 24 WHEN 'LU' THEN 20 WHEN 'DK' THEN 18 WHEN 'PL' THEN 28 WHEN 'CZ' THEN 24 WHEN 'GB' THEN 22 WHEN 'IE' THEN 22 WHEN 'SE' THEN 24 WHEN 'FI' THEN 18 WHEN 'NO' THEN 15 WHEN 'PT' THEN 25 WHEN 'LI' THEN 21 ELSE 0 END AS L", vR)
    vWant = vR
    if ((vWant > 0 & vLen != vWant) | vLen < 15 | vLen > 34)
        pResult = "WRONG_LENGTH"
        return 0
    endif
    vRest = vLen - 4
    vX = $concat(vNorm[5:vRest], vNorm[1:4])
    vM = 0
    vI = 1
    while (vI <= vLen)
        vC = vX[vI:1]
        vD = $scan("0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ", vC) - 1
        if (vD < 10)
            vM = vM * 10 + vD
        else
            vM = vM * 100 + vD
        endif
        while (vM >= 97)
            vM = vM - 97
        endwhile
        vI = vI + 1
    endwhile
    if (vM != 1)
        pResult = "WRONG_CHECKSUM"
    endif
    return 0
end
Enter fullscreen mode Exit fullscreen mode

Notes on the code:

  • vNorm[5:vRest] is a substring from position 5 with length vRest; vNorm[1:4] are the first four characters. Together they are step 2.
  • $scan returns the position of the character in the list 0-9A-Z, so position minus 1 is exactly the value from step 3: digits 0 to 9, letters 10 to 35.
  • The GLOB check runs first, so the loop only ever sees digits and capital letters.
  • The length table covers the countries this app deals with; for any other country only the general limit of 15 to 34 characters applies.
  • The length CASE has an alias (AS L). The companion article on debugging explains why that matters with this driver.

GET_TEXT is a small helper that runs a query and returns the first column of the first row:

entry GET_TEXT
params
    string pSql : IN
    string pValue : OUT
endparams
variables
    string vData, vRow
endvariables
    pValue = ""
    sql/data pSql, "CUSTOMERS"
    if ($status < 0)
        return -1
    endif
    vData = $result
    if (vData = "")
        return 0
    endif
    getitem vRow, vData, 1
    getitem pValue, vRow, 1
    return 0
end
Enter fullscreen mode Exit fullscreen mode

Trying it

Three public example IBANs and one with a changed last digit:

Input Result
DE89 3704 0044 0532 0130 00 OK
GB82 WEST 1234 5698 7654 32 OK
DE89 3704 0044 0532 0130 01 WRONG_CHECKSUM
AT61 1904 3002 3457 320 WRONG_LENGTH

The same results come from the automatic test service of the module, which also stores a seller IBAN through the form logic and expects the normalised value DE89370400440532013000 in the settings table.

Why not in SQL?

A first version computed the remainder with a recursive CTE and the % operator in SQLite. It is correct in the sqlite3 shell. When the test process of this module crashed inside this entry, the CTE was the first suspect; the real cause turned out to be another statement (described in the companion article). I kept the ProcScript loop anyway: 30 iterations, no driver involved, easy to step through in the debugger.

Limits

  • A valid check digit does not mean the account exists. It catches typos and swapped digits, which is what a data entry form needs.
  • The country length table is deliberately short. Extend it from the official IBAN registry if your users enter accounts from more countries.

Top comments (0)