Support deploying syndicated views to stage (#8532)

* Support for deploying syndicated views to stage

* Don't publish or replace references of syndicate views in stage deploys

* Address review feedback

* test app store view deploys

* get schemas for tables that aren't in the repo

* undo app store test

* Address review feedback

Author: scholtzan

bigquery_etl/cli/stage.py

@@ -310,8 +310,18 @@ def _view_dependencies(artifact_files, sql_dir):
                 if not path.exists():
                     path.mkdir(parents=True, exist_ok=True)
                     # don't create schema for wildcard and metadata tables
+                    # Create schemas for syndicated tables, stable, live tables and other
+                    # tables not managed by bigquery-etl by doing a dryrun.
+                    # The stage project doesn't have access to prod tables (e.g when referenced)
+                    # so we need to create the schema here and deploy it.
                     if "*" not in name and name != "INFORMATION_SCHEMA":
-                        partitioned_by = "submission_timestamp"
+                        partitioned_by = None
+
+                        if any(
+                            dataset.endswith(suffix) for suffix in ("_live", "_stable")
+                        ):
+                            partitioned_by = "submission_timestamp"
+
                         schema = Schema.for_table(
                             project=project,
                             dataset=dataset,
@@ -321,11 +331,11 @@ def _view_dependencies(artifact_files, sql_dir):
                         )
                         schema.to_yaml_file(path / SCHEMA_FILE)
 
-                if not file_exists_for_dependency:
-                    (path / QUERY_SCRIPT).write_text(
-                        "# Table stub generated by stage deploy"
-                    )
-                    view_dependencies.add(path / QUERY_SCRIPT)
+                    if not file_exists_for_dependency:
+                        (path / QUERY_SCRIPT).write_text(
+                            "# Table stub generated by stage deploy"
+                        )
+                        view_dependencies.add(path / QUERY_SCRIPT)
 
             # extract UDF references from view definition
             raw_routines = read_routine_dir()

bigquery_etl/view/__init__.py

@@ -426,9 +426,9 @@ def publish(self, target_project=None, dry_run=False, client=None):
 
                 if not self.metadata:
                     print(f"Missing metadata for {self.path}")
-
-                table.description = self.metadata.description
-                table.friendly_name = self.metadata.friendly_name
+                else:
+                    table.description = self.metadata.description
+                    table.friendly_name = self.metadata.friendly_name
 
                 if table.labels != self.labels:
                     labels = self.labels.copy()