BigQuery Mobile: Small changes to reduce errors during initial deployment

[aidanradford]

This looks like a copy paste error - please delete.

Typically our default customer schema name for the atomic dataset in bigquery is 'rt_pipeline_prod1', would it be possible to set this as the default here please?

[colmsnowplow]

This looks like a copy paste error - please delete.

Not a copy paste error. The stored procedures are used in each module, the step that creates them needs to be run for each module.

Typically our default customer schema name for the atomic dataset in bigquery is 'rt_pipeline_prod1', would it be possible to set this as the default here please?

This is an open source repo, and the name of the schema/dataset everywhere else across our estate of documentation and resources is 'atomic'. Up to @bill-warner to decide how to handle but IMO it doesn't make much sense to refer to the default that we use in our deployment here, since to understand what it is you need knowledge of our internal deployments.

It occurs in one module, it can't be that much effort to change it when deploying a model, surely? Especially when you should be double-checking the playbook values before running a model regardless.

[bill-warner]

Hi Aidan - thanks for the feedback. As Colm said, the loading of procedures in each module is indeed intentional. Since the procedures are dropped in the complete step of each module, and not all modules need to be executed during a run, we must load the procedures when we initialise each module.

Regarding the default schemas, I am with Colm here. I would rather keep consistency across all 3 versions of the model. I think the real problem here is the lack of global variables within sql-runner.

Please let me know if anything else comes up during deployment