I am not sure if this is supported. We are implementing the cx_Oracle library for all of our attribute queries and other data processing (it's then utilized with arcpy in one form or another). It is a bit of a pain to get the field mapping done, but it has been an overall great way to integrate non-spatial data stored in Oracle db's.
According to this 10.0 Help page, it seems to be returning either a List object in case of SQL SELECT statement, or a Boolean for fail/success when running a schema changing SQL query (DDL), or UPDATE / INSERT statement:
Thanks for the replies. I'm aware of the options you have suggested but I really need to invoke a function because one of the parameters needs to be a clob. I could use the cx_Oracle module but then I get back a non-ESRI cursor that I then have to convert. Ordinarily not a big deal but I have to do this a lot on a lot of data and in the interest of performance and maintenance. I see v.10.1 has expanded the arcpy.da module with functionality that may do the trick.
Thanks for the replies. I'm aware of the options you have suggested but I really need to invoke a function because one of the parameters needs to be a clob. I could use the cx_Oracle module but then I get back a non-ESRI cursor that I then have to convert. Ordinarily not a big deal but I have to do this a lot on a lot of data and in the interest of performance and maintenance. I see v.10.1 has expanded the arcpy.da module with functionality that may do the trick.
Just to throw this out there, what about NumPyArrayToTable? Wondering if you just add the cx_Oracle cursor (from the result of your function call) into an array then let the method do the work -- rather than doing all the gymnastics to get the field mapping done. Not sure but I may have to look into this myself.
