JPMML Classifier ERROR: SAX2 driver class org.apache.xerces.parsers.SAXParser not found

Dear KNIMErs,

I'm trying to use the node JPMML classifier to make prediction but I obtain the following error:

ERROR JPMML Classifier 0:1483 Execute failed: SAX2 driver class org.apache.xerces.parsers.SAXParser not found

I'm using KNIME 2.12.0 on a GNU/Linux OS.

Please, do you have any suggestion?

Gio

Hi Gio,

sorry, I can't really say what the problem is. On our Linux machine the node works. Can you tell me which Java version you are using?

Regards,

Alexander

Hi Alexander,

Thank you for your reply. I'm using KNIME version 2.12.1 running on a GNU/Linux (Ubuntu 14.04) OS. As far as I know KNIME is using the bundled Java version 1.7.0. Is there something I can do to solve this problem?

Best,

Gio

Hi,

I have no clue why this error occurs on your system, but maybe you can use one of the other predictors? We also have a PMML Predictor node and a Compiling PMML Predictor in KNIME Labs.

Regards,

Alexander

Hi,

I have the following error: ERROR JPMML Classifier 0:12 Execute failed: The target value given back by JPMML has the wrong type.

I dont understand, because i am using the same information with wich use for execute the model.

Please, help me..

Thanks

Hi guys,

I came back to this issue because I just realized that this problem appears and disappears (apparently randomly) from one KNIME session to another. This means that sometimes the JPMML classifier node gives the problems and cannot be run, and sometimes if you just restart KNIME, it correctly works. Is it possible that this depends on the order by which dependencies are loaded? Is there a way to avoid the problem?

Hi,

I'm experiencing the same issue in the latest KNIME release 3.40 with the examples on the knime server.

I'm using the latest Windows OS.

cheers,

mike

Hi,
thank you for reporting this. That it occurs kind of randomly really suggests that there is a problem with conflicts with other dependencies. I will have a chat with some other developers and see what their opinion is.
Kind regards,
Alexander