How to recode many fields in Kettle, all of which require the same recoding

Question

I have many fields all needing the same recoding, is there a transformation that will allow me to specify all the fields requiring the same mapping, rather that having to create a Value Mapper transformation for each field? I am using Kettle Spoon 4.4

Answer 1

I am not a huge fan of Java Scripting on Kettle, but I don't know of any other standard step that does what you wish.

You can add a Modified Java Script step to your stream and write a simple code to map the values of a list of fields (columns). To accomplish this, suppose you have fields A and B both with two possible values S for small and B for big. To map them you should insert the following javascript code:

// list of fields you wish to map
var fieldsToMap = ["A", "B"];
var tmpField;
var fieldIndex;

// for each field to map...
for (var i=0; i < fieldsToMap.length; i++) {
    //get the index of the field once you only have it's name
    fieldIndex = getInputRowMeta().indexOfValue(fieldsToMap[i]);
    //get the field
    tmpField = row.getValue(fieldIndex);
    //don't forget to trim as Kettle usually pads strings
    switch (trim(tmpField)) {
        case "S": 
            tmpField.setValue("Small"); 
            break;
        case "B": 
            tmpField.setValue("Big"); 
            break;
    }
}

Don't forget to check Compatibility mode? at the Java Script step, otherwise you'll have a javascript error once with the new standards you are not allowed to change a field value any more (see How to modify values (with compatibility off) at http://wiki.pentaho.com/display/EAI/Modified+Java+Script+Value)