Only join required tables

[philippthun]

ServicePlanListFetcher and ServiceOfferingListFetcher JOIN multiple tables to be used in WHERE conditions. The actual set of tables to be joined depends on the provided filters. Instead of joining a superset of tables, this change ensures that only the required tables will be joined.

Unfortunately Postgres does not remove unnecessary JOINs and so they possibly result in slower query execution times.

Before using another table in a WHERE clause, one simply has to call join_<table>(dataset). This can even be done multiple times for each table as the underlying join(dataset, type, table, on) function ensures that only unique table expressions are added.

• I have reviewed the contributing guide

• I have viewed, signed, and submitted the Contributor License Agreement

• I have made this pull request to the main branch

• I have run all the unit tests using bundle exec rake

• I have run CF Acceptance Tests

[philippthun]

We ran EXPLAIN (ANALYZE, BUFFERS) for a SELECT DISTINCT "service_plans".* FROM "service_plans" query with the following WHERE condition:

WHERE ((("service_plans"."public" IS TRUE)
  OR ("plan_orgs"."guid" IN ('org_guid'))
  OR ("broker_spaces"."guid" IN ('space_guid')))
AND ("service_instances"."guid" IN ('service_instance_guid')));

The main indicators were:

Cost: 324,264.94
Buffers: shared hit=67150, temp read=16048 written=16048
Execution Time: 3331.547 ms

After removing the (in this case) unnecessary joins LEFT JOIN "organizations" AS "broker_orgs" and LEFT JOIN "spaces" AS "plan_spaces" the EXPLAIN (ANALYZE, BUFFERS) changes to:

Cost=32,145.98
Buffers: shared hit=14007
Execution Time: 65.044 ms

So omitting unnecessary JOINs can have a huge impact (on Postgres)!