When having a seemly elaborate workbench, would run it using a specific data. Now when this same workbench is to be used for several different data... traces of the attributes of the earlier data exists.
Anyway to refresh the workbench such that these traces may be removed from each transformer ?
Thank You.
Own is right, but be aware that there are also some transformers that introduce attributes not directly related to a reader. Examples include the PythonCaller and the SQLExecutor, where you can enter a list of new attributes to expose. These will persist even though the readers are deleted.
And of course there are the AttributeCreators, ParameterFetchers, etc...
David
This is a problem i encounter often.
Lol 10 minutes ago i used a bench that reads a Citygml and extracts a set of data out of it.
I removed the old readers, an GML reader and a XML reader.
End inserted new ones pointing to the new dataset.
No dice! Constant errors. While the data schema is exactly the same.
After 3-4 tries and changing settings, i decided to make a fresh bench and enter the readers pointing to the new data...waddayaknow....it worked at the first attempt.
Purge helps sometimes if ur lucky.
It kinda reminds me of ARcmap .mxd behaviour, keeping a hold ondata. Unless you shutdown and restart the program.
Prior to 2013 ps 2 FME would sometimes suddenly mix the attributes..i would get like streetnames in housnumber attributes, housnumber in citynames etc. The only solution then wat to shut down all instances of fme, including viewers, inspectors.
Hav'nt seen that happen in fme 2013 sp2 yet.
(We will wait for fme 2014 sp+ before we use that in production.)