extract large object data types from oracle to celonis ems with jdbc extractor

Dear @markus.weih12 ,

You may want to check https://www.vertica.com/docs/12.0.x/HTML/Content/Authoring/SQLReferenceManual/DataTypes/OracleVerticaDataTypeMappings.htm?. You also may perform pre-process to ensure compatibility before extracting into Celonis.

Let me know your thoughts.

Thanks for your reply!

What does you mean with pre-processing?

Instead of perform conversion during Celonis extraction, perform all necessary clean up at source by creating another view or something similar and then extract into Celonis. This will make your script leaner / less to maintain.

Thanks. Is there any alternative option to convert during extraction on celonis side? In our setting it is not easily possible to change or add anything in source system.

Markus, sure you can do it during task creation under extraction job. At the task config, under General > Column Configuration then you may select the data type.

If extraction didnt provide the right setting and maybe try it as string / text first and later create another staging with the right conversion once data in celonis. Let me know and hopefully it works.

Unfortunately we receive only IDs like oracle.sql.CLOB@xxxxxxx within Celonis. So I think we have only the opportunity to create a view or something similar in source system and then extract into Celonis.

Source is better, stage it in Azure or other cloud also works it depend on overall architecture. It is interesting question, unfortunately I dont have Oracle db for me to setup quick trial to test some idea. Sorry for not being helpful.

Btw @Guillermo Gost  mention interesting area which is PyCelonis. Perhaps another area to explore to process it through ML workbench https://celonis.github.io/pycelonis/2.0.0/tutorials/executed/02_data_integration/02_data_push/. In short you may create your predefine table and then get data using panda + perform conversion and then push to table.

@Teuku Faruq thanks for your ideas!

Thanks @david.stefa13 - that is very cool and helps us!