Here is your corrected text:

I can do simple treatments in Pentaho, like reading a CSV file and filling the database, but I can’t find the right method to manage the resumption of the calculation of an identifier between 2 import sessions.

In a Pentaho treatment, I need to perform the following operations:

Read the lines of a CSV file. Fill an Oracle database with the lines from the CSV. To fill the Oracle database, I must fill in, for each line of the CSV, a field whose data must be unique. So I need a counter composed as follows: “ID_” + a sequence of the alphabet (A, B, C, etc.) + 2 digits. Examples: ID_A01, ID_A02, ID_A03 … ID_A99, ID_B01, ID_B02.

To calculate the ID, I use this Javascript code:

var lastSeq = LAST_SEQ;
var newSeq = "";

if (lastSeq === null || lastSeq === "") {
    newSeq = "ID_A01";
} else {
    var regex = /ID_([A-Z])(dd)/;
    var match = lastSeq.match(regex);
    if (match) {
        var letter = match[1];
        var number = parseInt(match[2]);
        number++;
        if (number > 99) {
            var newLetter = String.fromCharCode(letter.charCodeAt(0) + 1);
            number = 1;
            newSeq = "ID_" + newLetter + ("0" + number).slice(-2);
        } else {
            newSeq = "ID_" + letter + ("0" + number).slice(-2);
        }
    }
}

The sequence fills the IDFILE field in the Oracle FILE table.

Be careful, the sequence must continue between the different import sessions.

I can use this SQL query to find the previous value that was in FILE:

SELECT MAX(IDFILE) AS LAST_SEQ FROM FILE
WHERE IDFILE LIKE 'ID_%'
AND Rownum = 1

Thank you in advance for your advice.

I have tried different approaches (like a call to a sub-transformation) to do this treatment, but for the moment, I get nothing good.