Merge branch 'main' into b338873783-matrix-factorization

bigframes/pandas/__init__.py

@@ -42,6 +42,7 @@
     read_gbq,
     read_gbq_function,
     read_gbq_model,
+    read_gbq_object_table,
     read_gbq_query,
     read_gbq_table,
     read_json,
@@ -306,6 +307,7 @@ def reset_session():
     "read_gbq",
     "read_gbq_function",
     "read_gbq_model",
+    "read_gbq_object_table",
     "read_gbq_query",
     "read_gbq_table",
     "read_json",

bigframes/pandas/io/api.py

@@ -193,6 +193,21 @@ def read_gbq_model(model_name: str):
 read_gbq_model.__doc__ = inspect.getdoc(bigframes.session.Session.read_gbq_model)
 
 
+def read_gbq_object_table(
+    object_table: str, *, name: Optional[str] = None
+) -> bigframes.dataframe.DataFrame:
+    return global_session.with_default_session(
+        bigframes.session.Session.read_gbq_object_table,
+        object_table,
+        name=name,
+    )
+
+
+read_gbq_object_table.__doc__ = inspect.getdoc(
+    bigframes.session.Session.read_gbq_object_table
+)
+
+
 def read_gbq_query(
     query: str,
     *,

bigframes/session/__init__.py

@@ -1625,6 +1625,9 @@ def from_glob_path(
     ) -> dataframe.DataFrame:
         r"""Create a BigFrames DataFrame that contains a BigFrames Blob column from a global wildcard path.
 
+        .. note::
+            BigFrames Blob is still under experiments. It may not work and subject to change in the future.
+
         Args:
             path (str):
                 The wildcard global path, such as "gs://<bucket>/<folder>/\*".
@@ -1641,6 +1644,7 @@ def from_glob_path(
         if not bigframes.options.experiments.blob:
             raise NotImplementedError()
 
+        # TODO(garrettwu): switch to pseudocolumn when b/374988109 is done.
         connection = connection or self._bq_connection
         connection = bigframes.clients.resolve_full_bq_connection_name(
             connection,
@@ -1653,6 +1657,33 @@ def from_glob_path(
         s = self.read_gbq(table)["uri"].str.to_blob(connection)
         return s.rename(name).to_frame()
 
+    def read_gbq_object_table(
+        self, object_table: str, *, name: Optional[str] = None
+    ) -> dataframe.DataFrame:
+        """Read an existing object table to create a BigFrames Blob DataFrame. Use the connection of the object table for the connection of the blob.
+        This function dosen't retrieve the object table data. If you want to read the data, use read_gbq() instead.
+
+        .. note::
+            BigFrames Blob is still under experiments. It may not work and subject to change in the future.
+
+        Args:
+            object_table (str): name of the object table of form <PROJECT_ID>.<DATASET_ID>.<TABLE_ID>.
+            name (str or None): the returned blob column name.
+
+        Returns:
+            bigframes.pandas.DataFrame:
+                Result BigFrames DataFrame.
+        """
+        if not bigframes.options.experiments.blob:
+            raise NotImplementedError()
+
+        # TODO(garrettwu): switch to pseudocolumn when b/374988109 is done.
+        table = self.bqclient.get_table(object_table)
+        connection = table._properties["externalDataConfiguration"]["connectionId"]
+
+        s = self.read_gbq(object_table)["uri"].str.to_blob(connection)
+        return s.rename(name).to_frame()
+
 
 def connect(context: Optional[bigquery_options.BigQueryOptions] = None) -> Session:
     return Session(context)