Skip to main content

We recently upgraded from 2018 to 2019.1.3.1, a workbench that worked fine in 2018 is now have python issues both in Server and desktop. What could have changed and anyone know how to fix it?

 

 

2019-10-29 10:37:06| 46.1| 0.2|FATAL |AttributePivoter (PythonFactory): PythonFactory failed to close properly

2019-10-29 10:37:06| 46.1| 0.0|ERROR |BADNEWS: A fatal error has occurred. Check the logfile above for details auto __thiscall STFDynamicFactory::doneAllGroups::<lambda_faaa637947b31c6de3e253c564f4e0ca>::operator ()(void) const(c:\\code\\fme\\foundation\\pluginbuilder\\pluginbuilder_cpp\\dynafact.cpp:169) - auto __thiscall STFDynamicFactory::doneAllGroups::<lambda_faaa637947b31c6de3e253c564f4e0ca>::operator ()(void) const(c:\\code\\fme\\foundation\\pluginbuilder\\pluginbuilder_cpp\\dynafact.cpp:169)

2019-10-29 10:37:06| 46.1| 0.0|ERROR |AttributePivoter (PythonFactory): A fatal error has occurred. Check the logfile above for details

2019-10-29 10:37:07| 46.5| 0.0|ERROR |A fatal error has occurred. Check the logfile above for details

2019-10-29 10:37:07| 46.5| 0.0|ERROR |auto __thiscall STFDynamicFactory::doneAllGroups::<lambda_faaa637947b31c6de3e253c564f4e0ca>::operator ()(void) const(c:\\code\\fme\\foundation\\pluginbuilder\\pluginbuilder_cpp\\dynafact.cpp:169) - auto __thiscall STFDynamicFactory::doneAllGroups::<lambda_faaa637947b31c6de3e253c564f4e0ca>::operator ()(void) const(c:\\code\\fme\\foundation\\pluginbuilder\\pluginbuilder_cpp\\dynafact.cpp:169) - class StatusInfo *__thiscall STFDynamicFactory::doneAllGroups(void)(c:\\code\\fme\\foundation\\pluginbuilder\\pluginbuilder_cpp\\dynafact.cpp:173) - class StatusInfo *__thiscall STFGroupByFactory::allDone(void)(c:\\code\\fme\\foundation\\factories\\groupbyfactory.cpp:640) - class StatusInfo *__thiscall STFFactoryPipeline::allDone(void)(c:\\code\\fme\\foundation\\factories\\pipeline.cpp:1322) - class StatusInfo *__thiscall STFTranslator::abortProcessing(class StatusInfo *)(c:\\code\\fme\\foundation\\framework\\engine\\stftrans.cpp:323) - class StatusInfo *__thiscall STFSimpleTranslator::convert(void)(c:\\code\\fme\\foundation\\kernel\\simptran.cpp:758) - class StatusInfo *__thiscall STFTranslator::translate(const class FMETValOrderedVector<class ObsoleteString> &,const class FMETValOrderedVector<class ObsoleteString> &)(c:\\code\\fme\\foundation\\framework\\engine\\stftrans.cpp:293)

Have you tried upgrading that specific transformer? Or simply deleting and re-inserting the AttributePivoter in FME 2019?


Have you tried upgrading that specific transformer? Or simply deleting and re-inserting the AttributePivoter in FME 2019?

no option to upgrade so i replaced it. It gave me the same error.


no option to upgrade so i replaced it. It gave me the same error.

Sounds like you may need to involve Safe support on this one.


Sounds like you may need to involve Safe support on this one.

submitting a case now

 


no option to upgrade so i replaced it. It gave me the same error.

I turned on reading warnings with errors and i got additional info.

2019-10-29 12:08:59| 17.3| 0.3|WARN |Python Exception <TypeError>: '<' not supported between instances of 'str' and 'float'

 

This led me to believe it is a problem with the data. I filtered out the null values and it ran fine. So maybe with the upgrade python 2.7 was able to ignore null? the newer version doesn't?

 

I'm guessing! any thoughts

 


Sounds like it's a similar issue to this

https://knowledge.safe.com/questions/97469/statisticscalculator-error-with-python-36.html


yes this is it! i placed a null attribute mapper and replaced null with 0 and it worked.


I turned on reading warnings with errors and i got additional info.

2019-10-29 12:08:59| 17.3| 0.3|WARN |Python Exception <TypeError>: '<' not supported between instances of 'str' and 'float'

 

This led me to believe it is a problem with the data. I filtered out the null values and it ran fine. So maybe with the upgrade python 2.7 was able to ignore null? the newer version doesn't?

 

I'm guessing! any thoughts

 

I'm guessing that it's probably an issue with the code in the AttributePivoter that may not handle NULL vales. What happens if you use the NullAttributeMapper to map NULL attribute values to empty strings?


Reply