temp

jorisvandenbossche · Dec 7, 2023 · 9a6ebf2 · 9a6ebf2
1 parent 4205837
commit 9a6ebf2
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 1 deletion.
diff --git a/cpp/src/arrow/dataset/file_parquet_test.cc b/cpp/src/arrow/dataset/file_parquet_test.cc
@@ -661,6 +661,19 @@ TEST_P(TestParquetFileFormatScan, PredicatePushdownRowGroupFragments) {
                            less(field_ref(FieldRef("struct", "i32")), literal(6)));
   CountRowGroupsInFragment(fragment, {1},
                            equal(field_ref(FieldRef("struct", "str")), literal("2")));
+
+  // unsupported combination of field type and kernel gives an error
+  auto filter = less(field_ref("list"), literal(6));
+  auto parquet_fragment = checked_pointer_cast<ParquetFileFragment>(fragment);
+  Status filter_status = parquet_fragment->SplitByRowGroup(filter).status();
+  EXPECT_RAISES_WITH_MESSAGE_THAT(
+      NotImplemented,
+      testing::HasSubstr("Function 'less' has no kernel matching"),
+      filter_status);
+
+  // filter on unsupported type (list) will return all RowGroups
+  CountRowGroupsInFragment(fragment, all_row_groups, less(field_ref("list"), literal(6)));
+
 }
 
 TEST_P(TestParquetFileFormatScan, ExplicitRowGroupSelection) {

diff --git a/python/pyarrow/_dataset_parquet.pyx b/python/pyarrow/_dataset_parquet.pyx
@@ -471,7 +471,7 @@ cdef class ParquetFileFragment(FileFragment):
 
         if filter is not None:
             schema = schema or self.physical_schema
-            c_filter = _bind(filter, schema)
+            c_filter = filter.unwrap() #_bind(filter, schema)
             with nogil:
                 c_fragment = move(GetResultValue(
                     self.parquet_file_fragment.SubsetWithFilter(