ISSUE-60: Implement get_view_names()

Author: dimkonko

pybigquery/sqlalchemy_bigquery.py

@@ -3,6 +3,8 @@
 from __future__ import absolute_import
 from __future__ import unicode_literals
 
+import operator
+
 from google import auth
 from google.cloud import bigquery
 from google.cloud.bigquery import dbapi, QueryJobConfig
@@ -283,6 +285,11 @@ def __init__(
     def dbapi(cls):
         return dbapi
 
+    @staticmethod
+    def _build_formatted_table_id(table):
+        """Build '<dataset_id>.<table_id>' string using given table."""
+        return "{}.{}".format(table.reference.dataset_id, table.table_id)
+
     @staticmethod
     def _add_default_dataset_to_job_config(job_config, project_id, dataset_id):
         # If dataset_id is set, then we know the job_config isn't None
@@ -349,6 +356,26 @@ def _json_deserializer(self, row):
         """
         return row
 
+    def _get_table_or_view_names(self, connection, table_type, schema=None):
+        current_schema = schema or self.dataset_id
+        get_table_name = self._build_formatted_table_id \
+            if self.dataset_id is None else \
+            operator.attrgetter("table_id")
+
+        client = connection.connection._client
+        datasets = client.list_datasets()
+
+        result = []
+        for dataset in datasets:
+            if current_schema is not None and current_schema != dataset.dataset_id:
+                continue
+
+            tables = client.list_tables(dataset.reference)
+            for table in tables:
+                if table_type == table.table_type:
+                    result.append(get_table_name(table))
+        return result
+
     @staticmethod
     def _split_table_name(full_table_name):
         # Split full_table_name to get project, dataset and table name
@@ -464,23 +491,13 @@ def get_table_names(self, connection, schema=None, **kw):
         if isinstance(connection, Engine):
             connection = connection.connect()
 
-        datasets = connection.connection._client.list_datasets()
-        result = []
-        for d in datasets:
-            if schema is not None and d.dataset_id != schema:
-                continue
+        return self._get_table_or_view_names(connection, "TABLE", schema)
 
-            if self.dataset_id is not None and d.dataset_id != self.dataset_id:
-                continue
+    def get_view_names(self, connection, schema=None, **kw):
+        if isinstance(connection, Engine):
+            connection = connection.connect()
 
-            tables = connection.connection._client.list_tables(d.reference)
-            for t in tables:
-                if self.dataset_id is None:
-                    table_name = d.dataset_id + '.' + t.table_id
-                else:
-                    table_name = t.table_id
-                result.append(table_name)
-        return result
+        return self._get_table_or_view_names(connection, "VIEW", schema)
 
     def do_rollback(self, dbapi_connection):
         # BigQuery has no support for transactions.

scripts/load_test_data.sh

@@ -6,6 +6,7 @@ bq rm -f -t test_pybigquery.sample
 bq rm -f -t test_pybigquery_alt.sample_alt
 bq rm -f -t test_pybigquery.sample_one_row
 bq rm -f -t test_pybigquery.sample_dml
+bq rm -f -t test_pybigquery.sample_view
 bq rm -f -t test_pybigquery_location.sample_one_row
 
 bq mk --table --schema=$(dirname $0)/schema.json --time_partitioning_field timestamp --clustering_fields integer,string test_pybigquery.sample
@@ -17,3 +18,5 @@ bq load --source_format=NEWLINE_DELIMITED_JSON --schema=$(dirname $0)/schema.jso
 
 bq --location=asia-northeast1 load --source_format=NEWLINE_DELIMITED_JSON --schema=$(dirname $0)/schema.json test_pybigquery_location.sample_one_row $(dirname $0)/sample_one_row.json
 bq mk --schema=$(dirname $0)/schema.json -t test_pybigquery.sample_dml
+
+bq mk --use_legacy_sql=false --view 'SELECT string FROM test_pybigquery.sample' test_pybigquery.sample_view

test/test_sqlalchemy_bigquery.py

@@ -283,11 +283,13 @@ def test_tables_list(engine, engine_using_test_dataset):
     assert 'test_pybigquery.sample' in tables
     assert 'test_pybigquery.sample_one_row' in tables
     assert 'test_pybigquery.sample_dml' in tables
+    assert 'test_pybigquery.sample_view' not in tables
 
     tables = engine_using_test_dataset.table_names()
     assert 'sample' in tables
     assert 'sample_one_row' in tables
     assert 'sample_dml' in tables
+    assert 'sample_view' not in tables
 
 
 def test_group_by(session, table, session_using_test_dataset, table_using_test_dataset):
@@ -436,15 +438,27 @@ def test_table_names_in_schema(inspector, inspector_using_test_dataset):
     assert 'test_pybigquery.sample' in tables
     assert 'test_pybigquery.sample_one_row' in tables
     assert 'test_pybigquery.sample_dml' in tables
+    assert 'test_pybigquery.sample_view' not in tables
     assert len(tables) == 3
 
     tables = inspector_using_test_dataset.get_table_names()
     assert 'sample' in tables
     assert 'sample_one_row' in tables
     assert 'sample_dml' in tables
+    assert 'sample_view' not in tables
     assert len(tables) == 3
 
 
+def test_view_names(inspector, inspector_using_test_dataset):
+    view_names = inspector.get_view_names()
+    assert "test_pybigquery.sample_view" in view_names
+    assert "test_pybigquery.sample" not in view_names
+
+    view_names = inspector_using_test_dataset.get_view_names()
+    assert "sample_view" in view_names
+    assert "sample" not in view_names
+
+
 def test_get_indexes(inspector, inspector_using_test_dataset):
     for table in ['test_pybigquery.sample', 'test_pybigquery.sample_one_row']:
         indexes = inspector.get_indexes('test_pybigquery.sample')