Dataset Upsert

Jmoreno

I have been reading up on this and noticed that I in fact do NOT have the Upsert column enabled therefore according to the document I need to contact customer support. My question is, will I need to do this for every single current dataset and all future ones as well? Or is this a "flip the switch" kind of deal and it will be available for all?

GrantSmith

In workbench you define the upsert key in the job configuration. Are you wanting to enable upsert on a dataset which has duplicate upsert keys already? Each new job would need to have the upsert key defined as each dataset would have different keys.

Jmoreno

No I don't even see the Upsert column in the job configuration in the schema tab to be able to select my key. In the documentation it states that if it is not there to contact CS but my concern is will I need to do this every single time I create a new dataset

jaeW_at_Onyx

@Jmoreno

yes, you do have to contact support / your CSM to have upsert available as a feature in your instance.

by default the feature is not 'turned on.'

once enabled, you configure upsert on a case by case basis.

by default datasets are set up as REPLACE when you run a connector.

UPSERT is a variation of APPEND.

might sound tedious, but if you can imagine, domo has no way of knowing what the unique row identifier for a dataset is.

also, unless your dataset is in the hundreds of millions of rows, it typically doesn't make sense to use UPSERT. UPSERT just like normal SQL does not support soft delete, so if there's a chance that rows of data are deleted from your dataset, you'd have to make a second manual process to remove the deleted rows. (upsert by definition can only insert or update).