mozilla / bigquery-etl

Bigquery ETL

Home Page:https://mozilla.github.io/bigquery-etl

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Encode Derived Funnel Dimensions into UDF

data-sync-user opened this issue · comments

There are a number of dimensions (funnel_derived, partner_org, distribution_model, etc) that important for defining our funnels.

However, they’re currently being defined directly in a customer looker view. (for example, here)

It would be better to have them centrally located in a UDF definition, so that we have a code-defined source of truth for how these dimensions are defined.

Furthermore, we could make these available directly in the view version of the underlying tables.

┆Issue is synchronized with this Jira Story

➤ Su-Young Hong commented:

Currently documenting the dimensions and their definitions in this document:

Funnel Derived Dimensions UDF Specs ( https://docs.google.com/document/d/1jN4s6NUIrFvaPQ19VlhtAjR1F2MotJrl01HLoSvzIHw/edit#heading=h.bq4s3ikux802 ) (in progress)

Will update ticket when the document finished and ready for translation into code.

➤ Su-Young Hong commented:

Funnel Derived Dimensions UDF Specs ( https://docs.google.com/document/d/1jN4s6NUIrFvaPQ19VlhtAjR1F2MotJrl01HLoSvzIHw/edit )This spec sheet has been completed, and definitions logs. It’s ready for translation into UDFs + views now!

➤ Katie Windau commented:

Hello Su-Young Hong - This request has been done. I created the 9 new UDFs (I also did a little cleaning up/simplifying of the logic and added some test cases too).

The 9 new UDFs are as follows:

  1. mozdata.udf.distribution_model_clients
  2. mozdata.udf.distribution_model_ga_metrics
  3. mozdata.udf.distribution_model_installs
  4. mozdata.udf.funnel_derived_clients
  5. mozdata.udf.funnel_derived_ga_metrics
  6. mozdata.udf.funnel_derived_installs
  7. mozdata.udf.partner_org_clients
  8. mozdata.udf.partner_org_ga_metrics
  9. mozdata.udf.partner_org_installs

The PR that was merged on 4/5 is here: https://github.com/mozilla/private-bigquery-etl/pull/371 ( https://github.com/mozilla/private-bigquery-etl/pull/371|smart-link )

The UDFs should show up in prod in about a day usually, let me know if you have any questions or need anything else, thanks!