Looker typically generates these suggestions by executing a
SELECT DISTINCT query on the field. For some large tables this query can be too slow or create too large of a database load.
suggest_dimension you can make Looker query an alternative dimension for the suggestion values. If that dimension lives in a different explore, you can use
suggest_explore as well. Otherwise, it is not required.
In this example, instead of looking through a huge list of user IDs from the event table, we’ve told Looker to query the IDs from a user table instead:
In this example, the suggestion uses a dimension in another view joined to the explore: