Feat: Adds foreign_type_info attribute to table class and adds unit tests. (#2126)

chalmerlowe · web-flow · commit 2c1968115bef · 2025-02-21T12:45:15.000-05:00
* adds foreign_type_info attribute to table

* feat: Adds foreign_type_info attribute and tests

* updates docstrings for foreign_type_info

* Updates property handling, especially as regards set/get_sub_prop

* Removes extraneous comments and debug expressions

* Refactors build_resource_from_properties w get/set_sub_prop

* updates to foreign_type_info, tests and wiring

* Adds logic to detect non-Sequence schema.fields value

* updates assorted tests and logic
diff --git a/google/cloud/bigquery/_helpers.py b/google/cloud/bigquery/_helpers.py
@@ -978,11 +978,11 @@ def _build_resource_from_properties(obj, filter_fields):
     """
     partial = {}
     for filter_field in filter_fields:
-        api_field = obj._PROPERTY_TO_API_FIELD.get(filter_field)
+        api_field = _get_sub_prop(obj._PROPERTY_TO_API_FIELD, filter_field)
         if api_field is None and filter_field not in obj._properties:
             raise ValueError("No property %s" % filter_field)
         elif api_field is not None:
-            partial[api_field] = obj._properties.get(api_field)
+            _set_sub_prop(partial, api_field, _get_sub_prop(obj._properties, api_field))
         else:
             # allows properties that are not defined in the library
             # and properties that have the same name as API resource key
diff --git a/google/cloud/bigquery/schema.py b/google/cloud/bigquery/schema.py
@@ -15,10 +15,9 @@
 """Schemas for BigQuery tables / queries."""
 
 from __future__ import annotations
-import collections
 import enum
 import typing
-from typing import Any, cast, Dict, Iterable, Optional, Union
+from typing import Any, cast, Dict, Iterable, Optional, Union, Sequence
 
 from google.cloud.bigquery import _helpers
 from google.cloud.bigquery import standard_sql
@@ -489,6 +488,8 @@ def _parse_schema_resource(info):
         Optional[Sequence[google.cloud.bigquery.schema.SchemaField`]:
             A list of parsed fields, or ``None`` if no "fields" key found.
     """
+    if isinstance(info, list):
+        return [SchemaField.from_api_repr(f) for f in info]
     return [SchemaField.from_api_repr(f) for f in info.get("fields", ())]
 
 
@@ -501,40 +502,46 @@ def _build_schema_resource(fields):
     Returns:
         Sequence[Dict]: Mappings describing the schema of the supplied fields.
     """
-    return [field.to_api_repr() for field in fields]
+    if isinstance(fields, Sequence):
+        # Input is a Sequence (e.g. a list): Process and return a list of SchemaFields
+        return [field.to_api_repr() for field in fields]
+
+    else:
+        raise TypeError("Schema must be a Sequence (e.g. a list) or None.")
 
 
 def _to_schema_fields(schema):
-    """Coerce `schema` to a list of schema field instances.
+    """Coerces schema to a list of SchemaField instances while
+    preserving the original structure as much as possible.
 
     Args:
-        schema(Sequence[Union[ \
-            :class:`~google.cloud.bigquery.schema.SchemaField`, \
-            Mapping[str, Any] \
-        ]]):
-            Table schema to convert. If some items are passed as mappings,
-            their content must be compatible with
-            :meth:`~google.cloud.bigquery.schema.SchemaField.from_api_repr`.
+        schema (Sequence[Union[ \
+                   :class:`~google.cloud.bigquery.schema.SchemaField`, \
+                   Mapping[str, Any] \
+                       ]
+                   ]
+               )::
+            Table schema to convert. Can be a list of SchemaField
+            objects or mappings.
 
     Returns:
-        Sequence[:class:`~google.cloud.bigquery.schema.SchemaField`]
+        A list of SchemaField objects.
 
     Raises:
-        Exception: If ``schema`` is not a sequence, or if any item in the
-        sequence is not a :class:`~google.cloud.bigquery.schema.SchemaField`
-        instance or a compatible mapping representation of the field.
+        TypeError: If schema is not a Sequence.
     """
-    for field in schema:
-        if not isinstance(field, (SchemaField, collections.abc.Mapping)):
-            raise ValueError(
-                "Schema items must either be fields or compatible "
-                "mapping representations."
-            )
 
-    return [
-        field if isinstance(field, SchemaField) else SchemaField.from_api_repr(field)
-        for field in schema
-    ]
+    if isinstance(schema, Sequence):
+        # Input is a Sequence (e.g. a list): Process and return a list of SchemaFields
+        return [
+            field
+            if isinstance(field, SchemaField)
+            else SchemaField.from_api_repr(field)
+            for field in schema
+        ]
+
+    else:
+        raise TypeError("Schema must be a Sequence (e.g. a list) or None.")
 
 
 class PolicyTagList(object):
diff --git a/google/cloud/bigquery/table.py b/google/cloud/bigquery/table.py
@@ -21,7 +21,8 @@
 import functools
 import operator
 import typing
-from typing import Any, Dict, Iterable, Iterator, List, Optional, Tuple, Union
+from typing import Any, Dict, Iterable, Iterator, List, Optional, Tuple, Union, Sequence
+
 import warnings
 
 try:
@@ -66,6 +67,7 @@
 from google.cloud.bigquery.encryption_configuration import EncryptionConfiguration
 from google.cloud.bigquery.enums import DefaultPandasDTypes
 from google.cloud.bigquery.external_config import ExternalConfig
+from google.cloud.bigquery import schema as _schema
 from google.cloud.bigquery.schema import _build_schema_resource
 from google.cloud.bigquery.schema import _parse_schema_resource
 from google.cloud.bigquery.schema import _to_schema_fields
@@ -398,7 +400,7 @@ class Table(_TableBase):
         "partitioning_type": "timePartitioning",
         "range_partitioning": "rangePartitioning",
         "time_partitioning": "timePartitioning",
-        "schema": "schema",
+        "schema": ["schema", "fields"],
         "snapshot_definition": "snapshotDefinition",
         "clone_definition": "cloneDefinition",
         "streaming_buffer": "streamingBuffer",
@@ -411,6 +413,7 @@ class Table(_TableBase):
         "max_staleness": "maxStaleness",
         "resource_tags": "resourceTags",
         "external_catalog_table_options": "externalCatalogTableOptions",
+        "foreign_type_info": ["schema", "foreignTypeInfo"],
     }
 
     def __init__(self, table_ref, schema=None) -> None:
@@ -451,8 +454,20 @@ def schema(self):
                 If ``schema`` is not a sequence, or if any item in the sequence
                 is not a :class:`~google.cloud.bigquery.schema.SchemaField`
                 instance or a compatible mapping representation of the field.
+
+        .. Note::
+            If you are referencing a schema for an external catalog table such
+            as a Hive table, it will also be necessary to populate the foreign_type_info
+            attribute. This is not necessary if defining the schema for a BigQuery table.
+
+            For details, see:
+            https://cloud.google.com/bigquery/docs/external-tables
+            https://cloud.google.com/bigquery/docs/datasets-intro#external_datasets
+
         """
-        prop = self._properties.get(self._PROPERTY_TO_API_FIELD["schema"])
+        prop = _helpers._get_sub_prop(
+            self._properties, self._PROPERTY_TO_API_FIELD["schema"]
+        )
         if not prop:
             return []
         else:
@@ -463,10 +478,21 @@ def schema(self, value):
         api_field = self._PROPERTY_TO_API_FIELD["schema"]
 
         if value is None:
-            self._properties[api_field] = None
-        else:
+            _helpers._set_sub_prop(
+                self._properties,
+                api_field,
+                None,
+            )
+        elif isinstance(value, Sequence):
             value = _to_schema_fields(value)
-            self._properties[api_field] = {"fields": _build_schema_resource(value)}
+            value = _build_schema_resource(value)
+            _helpers._set_sub_prop(
+                self._properties,
+                api_field,
+                value,
+            )
+        else:
+            raise TypeError("Schema must be a Sequence (e.g. a list) or None.")
 
     @property
     def labels(self):
@@ -1075,6 +1101,43 @@ def external_catalog_table_options(
                 self._PROPERTY_TO_API_FIELD["external_catalog_table_options"]
             ] = value
 
+    @property
+    def foreign_type_info(self) -> Optional[_schema.ForeignTypeInfo]:
+        """Optional. Specifies metadata of the foreign data type definition in
+        field schema (TableFieldSchema.foreign_type_definition).
+
+        Returns:
+            Optional[schema.ForeignTypeInfo]:
+                Foreign type information, or :data:`None` if not set.
+
+        .. Note::
+            foreign_type_info is only required if you are referencing an
+            external catalog such as a Hive table.
+            For details, see:
+            https://cloud.google.com/bigquery/docs/external-tables
+            https://cloud.google.com/bigquery/docs/datasets-intro#external_datasets
+        """
+
+        prop = _helpers._get_sub_prop(
+            self._properties, self._PROPERTY_TO_API_FIELD["foreign_type_info"]
+        )
+        if prop is not None:
+            return _schema.ForeignTypeInfo.from_api_repr(prop)
+        return None
+
+    @foreign_type_info.setter
+    def foreign_type_info(self, value: Union[_schema.ForeignTypeInfo, dict, None]):
+        value = _helpers._isinstance_or_raise(
+            value,
+            (_schema.ForeignTypeInfo, dict),
+            none_allowed=True,
+        )
+        if isinstance(value, _schema.ForeignTypeInfo):
+            value = value.to_api_repr()
+        _helpers._set_sub_prop(
+            self._properties, self._PROPERTY_TO_API_FIELD["foreign_type_info"], value
+        )
+
     @classmethod
     def from_string(cls, full_table_id: str) -> "Table":
         """Construct a table from fully-qualified table ID.
diff --git a/tests/unit/job/test_load.py b/tests/unit/job/test_load.py
@@ -272,7 +272,7 @@ def test_schema_setter_invalid_field(self):
 
         config = LoadJobConfig()
         full_name = SchemaField("full_name", "STRING", mode="REQUIRED")
-        with self.assertRaises(ValueError):
+        with self.assertRaises(TypeError):
             config.schema = [full_name, object()]
 
     def test_schema_setter(self):
diff --git a/tests/unit/test_client.py b/tests/unit/test_client.py
@@ -2051,7 +2051,7 @@ def test_update_dataset(self):
         ds.labels = LABELS
         ds.access_entries = [AccessEntry("OWNER", "userByEmail", "phred@example.com")]
         ds.resource_tags = RESOURCE_TAGS
-        fields = [
+        filter_fields = [
             "description",
             "friendly_name",
             "location",
@@ -2065,12 +2065,12 @@ def test_update_dataset(self):
         ) as final_attributes:
             ds2 = client.update_dataset(
                 ds,
-                fields=fields,
+                fields=filter_fields,
                 timeout=7.5,
             )
 
         final_attributes.assert_called_once_with(
-            {"path": "/%s" % PATH, "fields": fields}, client, None
+            {"path": "/%s" % PATH, "fields": filter_fields}, client, None
         )
 
         conn.api_request.assert_called_once_with(
@@ -2615,7 +2615,7 @@ def test_update_table_w_schema_None(self):
         self.assertEqual(len(conn.api_request.call_args_list), 2)
         req = conn.api_request.call_args_list[1]
         self.assertEqual(req[1]["method"], "PATCH")
-        sent = {"schema": None}
+        sent = {"schema": {"fields": None}}
         self.assertEqual(req[1]["data"], sent)
         self.assertEqual(req[1]["path"], "/%s" % path)
         self.assertEqual(len(updated_table.schema), 0)
diff --git a/tests/unit/test_schema.py b/tests/unit/test_schema.py
diff --git a/tests/unit/test_table.py b/tests/unit/test_table.py
