Exact use of staging tables?

ravireddy25 · Post by **ravireddy25** » Thu Jun 04, 2009 1:36 am

arvind_ds · Post by **arvind_ds** » Thu Jun 04, 2009 1:58 am

Staging tables are used to store extracted data temporarily to avoid dependency on source systems. Once data is loaded into staging tables, it can further be loaded into desired dimension/detail/transaction/fact tables.
Its a kind of intermediate storage.

chulett · Post by **chulett** » Thu Jun 04, 2009 6:09 am

While I normally avoid crap sites like 'Geek Interview' there's a decent answer out there on it here. Typical nonsense here. In both cases the comments are fairly useless. Google up your own answers for more fun.

ray.wurlod · Post by **ray.wurlod** » Thu Jun 04, 2009 4:10 pm

Exact use of staging tables? Staging.

chulett · Post by **chulett** » Thu Jun 04, 2009 5:00 pm

That made me L-O-L.

dsex100 · Post by **dsex100** » Thu Jun 04, 2009 10:42 pm

They're also good to make sure that all your data has been loaded into the target data warehouse area (i.e. the "staging" area).

In the case where for example, not all your source data files where available then you could wait until all the staging jobs completed, at this point you'd know all your data is ready to be loaded into the data warehouse.

Then you can load your work tables using some transformations, the whole process happening in it's own separate environment.

ray.wurlod · Post by **ray.wurlod** » Thu Jun 04, 2009 11:53 pm

I prefer Data Sets for that (with parallel jobs). They are more efficiently loaded and unloaded than database tables.

Things like referential integrity should have been checked earlier in the ETL process - you don't need a "staging database" to do that.