Create test data for query integration tests
Currently, the query integration tests are very slow. They can also fail depending on the load on the database.
We can create a sample database that has all data that is needed for integration tests to pass. One way to create such a database is:
We can create unique identifiers (like id or hash, etc) of the data that is queried or returned by integration tests. Then a script should check dependencies of data (like, a bigmap depends on operation, account and block) and then calculate all rows that need to be loaded.
We need to maintain this code as new versions of indexer db are released.
This can make the integration tests very fast, while we can use the full database for performance tests.