I posted a similar question, but that one is not clear enough
I fellowed the 02_document_classfication example workflow.
The example workflow shows me how to test the model, now I got how to train a model, but how to use a trained SVM model to classify the ROW TEXT.
I am not sure what do you need. In the example workflow it does show how to classify the rows using an SVM model: using the SVM Predictor. You have to do the exactly the same transformations to the test dataset as to the training dataset and apply the model to the preprocessed test set by the SVM Predictor. The 02_document_classification workflow "cheats" a bit, because it transforms both together and only before learning splits them, but the general idea is the same. Maybe with PMML transformations you would only need to combine those to apply them on the test dataset with the SVM model, but not sure everything is available there (PMML Transform Applier might help and also this blogpost).