From 2dabf6cb22933fefa7daab1ade7076588f8c8756 Mon Sep 17 00:00:00 2001 From: lucia-vargas-a Date: Tue, 18 Jun 2024 21:22:04 +0200 Subject: [PATCH] Updates. --- src/cookbooks/data_modeling/using_aggregates.md | 1 + 1 file changed, 1 insertion(+) diff --git a/src/cookbooks/data_modeling/using_aggregates.md b/src/cookbooks/data_modeling/using_aggregates.md index 7950cc00d..f3f6fef07 100644 --- a/src/cookbooks/data_modeling/using_aggregates.md +++ b/src/cookbooks/data_modeling/using_aggregates.md @@ -63,6 +63,7 @@ This doc is about when to use different options to aggregate data, their limitat - All types of aggregates require a backfill when the source data changes: - BigQuery aggregate tables are backfilled using the managed-backfill process. - Materialized views cannot be backfilled, instead a materialized view needs to be re-created. Schema changes in base tables invalidate the view and requires it to be re-created. Materialized views scan all historical data of their referenced base tables by default, so **ensure to set a date filter to reduce the amount of data to be scanned**. + - Add a date filter to materialized views to limit the amount of data to be scanned when these views get deployed initially. Otherwise they will scan the entire data in referenced base tables. - Looker PDTs require following a [protocol to backfill](https://mozilla.udemy.com/course/looker-training-for-developers/learn/lecture/35440216/). - Using indices, partition and clustering is allowed for all cases. Looker requires that they are defined in the base table. - There is a limit of 20 materialized views per table.