Flattening the export means converting the nested event_params and user_properties arrays into top-level columns, so analysts can query events like a standard star schema. The common pattern is a dbt model that pivots known parameter keys with conditional aggregation, producing one column per parameter.
The GA4 BigQuery export is a native integration that streams or batches your raw event data into a BigQuery dataset, with one daily table named events_YYYYMMDD. Each row is a single event with nested user_properties and event_params arrays, giving you full SQL access to data that the GA4 UI aggregates and samples.
The GA4 BigQuery export schema stores each event as a row with nested REPEATED RECORD fields for event_params, user_properties, and items. To read a parameter value, you UNNEST the array and filter by key, which is verbose but gives you full access to every dimension GA4 captures.