0

I currently need to blend two data sources (two distinct BigQuery tables) in Google Data Studio which means every time a report filter selection is updated, Data Studio needs to re-compute the left outer join on the fly (or at least it looks as if it is because it takes ages to load). I was reading about BI Engine and wondered if that could be useful for this use case in speeding up left outer joins in Data Studio?

https://cloud.google.com/bi-engine/docs/getting-started-data-studio

mrag
  • 51
  • 3
  • 10
  • 1
    Consider pre-joining your tables into a new table using BQ scheduled queries. This way you only do the join once. Client applications (Data Studio in this case) will have simpler queries to run that don't include joins at 'run-time'. This process is called denormalization. For 2 tables, BI Engine is probably overkill. – rtenha Jan 29 '20 at 19:45
  • Thanks @rtenha. Any idea how to use Scheduled Query to join individual daily tables in a wildcard table with a second table into a new individual daily table in a second wildcard table? So, let's call tables in question table1, table2 and table3. table1 = original wildcard table with nested daily tables table2 = dynamically updating table every day table3 = resulting wildcard table from the join with nested daily tables How to left join table2 onto table1_YYYYMMDD and write output to table3_YYYYMMDD? – mrag Jan 30 '20 at 12:07

1 Answers1

1

Yes, it should be.

For this use case, you should create a new view in BigQuery that joins the two tables. Then query that view directly from Data Studio. The guide you linked to has information on how to enable BigQuery BI Engine for your GCP project.

Minhaz Kazi
  • 3,115
  • 1
  • 10
  • 20