Creating tables with a dask-cudf DataFrame

dask_sql/context.py

@@ -125,6 +125,7 @@ def create_table(
         format: str = None,
         persist: bool = True,
         schema_name: str = None,
+        gpu: bool = False,
         **kwargs,
     ):
         """
@@ -199,6 +200,7 @@ def create_table(
             table_name=table_name,
             format=format,
             persist=persist,
+            gpu=gpu,
             **kwargs,
         )
         self.schema[schema_name].tables[table_name.lower()] = dc

dask_sql/input_utils/convert.py

@@ -37,6 +37,7 @@ def to_dc(
         table_name: str,
         format: str = None,
         persist: bool = True,
+        gpu: bool = False,
         **kwargs,
     ) -> DataContainer:
         """
@@ -45,7 +46,7 @@ def to_dc(
         maybe persist them to cluster memory before.
         """
         filled_get_dask_dataframe = lambda *args: cls._get_dask_dataframe(
-            *args, table_name=table_name, format=format, **kwargs,
+            *args, table_name=table_name, format=format, gpu=gpu, **kwargs,
         )
 
         if isinstance(input_item, list):
@@ -60,7 +61,12 @@ def to_dc(
 
     @classmethod
     def _get_dask_dataframe(
-        cls, input_item: InputType, table_name: str, format: str = None, **kwargs,
+        cls,
+        input_item: InputType,
+        table_name: str,
+        format: str = None,
+        gpu: bool = False,
+        **kwargs,
     ):
         plugin_list = cls.get_plugins()
 
@@ -69,7 +75,7 @@ def _get_dask_dataframe(
                 input_item, table_name=table_name, format=format, **kwargs
             ):
                 return plugin.to_dc(
-                    input_item, table_name=table_name, format=format, **kwargs
+                    input_item, table_name=table_name, format=format, gpu=gpu, **kwargs
                 )
 
         raise ValueError(f"Do not understand the input type {type(input_item)}")

dask_sql/input_utils/intake.py

@@ -18,11 +18,21 @@ def is_correct_input(
             isinstance(input_item, intake.catalog.Catalog) or format == "intake"
         )
 
-    def to_dc(self, input_item: Any, table_name: str, format: str = None, **kwargs):
+    def to_dc(
+        self,
+        input_item: Any,
+        table_name: str,
+        format: str = None,
+        gpu: bool = False,
+        **kwargs
+    ):
         table_name = kwargs.pop("intake_table_name", table_name)
         catalog_kwargs = kwargs.pop("catalog_kwargs", {})
 
         if isinstance(input_item, str):
             input_item = intake.open_catalog(input_item, **catalog_kwargs)
 
-        return input_item[table_name].to_dask(**kwargs)
+        if gpu:
+            raise Exception("Intake does not support gpu")
+        else:
+            return input_item[table_name].to_dask(**kwargs)

dask_sql/input_utils/location.py

@@ -15,7 +15,14 @@ def is_correct_input(
     ):
         return isinstance(input_item, str)
 
-    def to_dc(self, input_item: Any, table_name: str, format: str = None, **kwargs):
+    def to_dc(
+        self,
+        input_item: Any,
+        table_name: str,
+        format: str = None,
+        gpu: bool = False,
+        **kwargs,
+    ):
 
         if format == "memory":
             client = default_client()
@@ -27,7 +34,12 @@ def to_dc(self, input_item: Any, table_name: str, format: str = None, **kwargs):
             format = extension.lstrip(".")
 
         try:
-            read_function = getattr(dd, f"read_{format}")
+            if gpu:
+                import dask_cudf
+
+                read_function = getattr(dask_cudf, f"read_{format}")
+            else:
+                read_function = getattr(dd, f"read_{format}")
         except AttributeError:
             raise AttributeError(f"Can not read files of format {format}")
 

dask_sql/input_utils/pandas.py

@@ -12,6 +12,21 @@ def is_correct_input(
     ):
         return isinstance(input_item, pd.DataFrame) or format == "dask"
 
-    def to_dc(self, input_item, table_name: str, format: str = None, **kwargs):
+    def to_dc(
+        self,
+        input_item,
+        table_name: str,
+        format: str = None,
+        gpu: bool = False,
+        **kwargs,
+    ):
         npartitions = kwargs.pop("npartitions", 1)
-        return dd.from_pandas(input_item, npartitions=npartitions, **kwargs)
+        if gpu:
+            import cudf
+            import dask_cudf
+
+            return dask_cudf.from_cudf(
+                cudf.from_pandas(input_item), npartitions=npartitions, **kwargs,
+            )
+        else:
+            return dd.from_pandas(input_item, npartitions=npartitions, **kwargs)

dask_sql/physical/rel/custom/create_table.py

@@ -62,11 +62,13 @@ def convert(
         except KeyError:
             raise AttributeError("Parameters must include a 'location' parameter.")
 
+        gpu = kwargs.pop("gpu", False)
         context.create_table(
             table_name,
             location,
             format=format,
             persist=persist,
             schema_name=schema_name,
+            gpu=gpu,
             **kwargs,
         )