diff --git a/bigframes/core/__init__.py b/bigframes/core/__init__.py
index 63f36d4ddd..b640692bc8 100644
--- a/bigframes/core/__init__.py
+++ b/bigframes/core/__init__.py
@@ -165,7 +165,7 @@ def cached(self, cluster_cols: typing.Sequence[str]) -> ArrayValue:
         ibis_expr = compiled_value._to_ibis_expr(
             ordering_mode="unordered", expose_hidden_cols=True
         )
-        tmp_table = self.session._ibis_to_session_table(
+        tmp_table = self.session._ibis_to_temp_table(
             ibis_expr, cluster_cols=cluster_cols, api_name="cached"
         )
 
diff --git a/bigframes/session/__init__.py b/bigframes/session/__init__.py
index bd5845631b..9b881de9a0 100644
--- a/bigframes/session/__init__.py
+++ b/bigframes/session/__init__.py
@@ -36,7 +36,6 @@
     Tuple,
     Union,
 )
-import uuid
 import warnings
 
 import google.api_core.client_info
@@ -836,11 +835,20 @@ def _read_bigquery_load_job(
             )
 
         self._start_generic_job(load_job)
+        table_id = f"{table.project}.{table.dataset_id}.{table.table_id}"
+
+        # Update the table expiration so we aren't limited to the default 24
+        # hours of the anonymous dataset.
+        table_expiration = bigquery.Table(table_id)
+        table_expiration.expires = (
+            datetime.datetime.now(datetime.timezone.utc) + constants.DEFAULT_EXPIRATION
+        )
+        self.bqclient.update_table(table_expiration, ["expires"])
 
         # The BigQuery REST API for tables.get doesn't take a session ID, so we
         # can't get the schema for a temp table that way.
         return self.read_gbq_table(
-            f"{table.project}.{table.dataset_id}.{table.table_id}",
+            table_id,
             index_col=index_col,
             col_order=col_order,
         )
@@ -977,7 +985,7 @@ def _read_pandas(
         job_config.clustering_fields = cluster_cols
         job_config.labels = {"bigframes-api": api_name}
 
-        load_table_destination = self._create_session_table()
+        load_table_destination = bigframes_io.random_table(self._anonymous_dataset)
         load_job = self.bqclient.load_table_from_dataframe(
             pandas_dataframe_copy,
             load_table_destination,
@@ -990,8 +998,9 @@ def _read_pandas(
             total_ordering_columns=frozenset([ordering_col]),
             integer_encoding=IntegerEncoding(True, is_sequential=True),
         )
-        table_expression = self.ibis_client.sql(
-            f"SELECT * FROM `{load_table_destination.table_id}`"
+        table_expression = self.ibis_client.table(
+            load_table_destination.table_id,
+            database=f"{load_table_destination.project}.{load_table_destination.dataset_id}",
         )
 
         # b/297590178 Potentially a bug in bqclient.load_table_from_dataframe(), that only when the DF is empty, the index columns disappear in table_expression.
@@ -1269,13 +1278,6 @@ def _check_file_size(self, filepath: str):
                 "for large files to avoid loading the file into local memory."
             )
 
-    def _create_session_table(self) -> bigquery.TableReference:
-        table_name = f"{uuid.uuid4().hex}"
-        dataset = bigquery.Dataset(
-            bigquery.DatasetReference(self.bqclient.project, "_SESSION")
-        )
-        return dataset.table(table_name)
-
     def _create_empty_temp_table(
         self,
         schema: Iterable[bigquery.SchemaField],
@@ -1310,7 +1312,7 @@ def _create_sequential_ordering(
             ibis.row_number().cast(ibis_dtypes.int64).name(default_ordering_name)
         )
         table = table.mutate(**{default_ordering_name: default_ordering_col})
-        table_ref = self._ibis_to_session_table(
+        table_ref = self._ibis_to_temp_table(
             table,
             cluster_cols=list(index_cols) + [default_ordering_name],
             api_name=api_name,
@@ -1326,7 +1328,7 @@ def _create_sequential_ordering(
         )
         return table, ordering
 
-    def _ibis_to_session_table(
+    def _ibis_to_temp_table(
         self,
         table: ibis_types.Table,
         cluster_cols: Iterable[str],