ignoring columns with unsupported parquet and json types with type inferring (#9524)

Author: evanevanevanevannnn

ydb/core/external_sources/object_storage/inference/arrow_inferencinator.cpp

@@ -126,32 +126,10 @@ bool ArrowToYdbType(Ydb::Type& maybeOptionalType, const arrow::DataType& type, s
     case arrow::Type::LIST: { // TODO: is ok?
         return false;
     }
-    case arrow::Type::STRUCT: { // TODO: is ok?
-        auto& structType = *resType.mutable_struct_type();
-        for (const auto& field : type.fields()) {
-            auto& member = *structType.add_members();
-            auto& memberType = *member.mutable_type();
-            if (!ArrowToYdbType(memberType, *field->type(), config)) {
-                return false;
-            }
-            member.mutable_name()->assign(field->name().data(), field->name().size());
-        }
-        return true;
-    }
+    case arrow::Type::STRUCT:
     case arrow::Type::SPARSE_UNION:
-    case arrow::Type::DENSE_UNION: { // TODO: is ok?
-        auto& variant = *resType.mutable_variant_type()->mutable_struct_items();
-        for (const auto& field : type.fields()) {
-            auto& member = *variant.add_members();
-            if (!ArrowToYdbType(*member.mutable_type(), *field->type(), config)) {
-                return false;
-            }
-            if (field->name().empty()) {
-                return false;
-            }
-            member.mutable_name()->assign(field->name().data(), field->name().size());
-        }
-        return true;
+    case arrow::Type::DENSE_UNION: {
+        return false;
     }
     case arrow::Type::DICTIONARY: // TODO: is representable?
         return false;
@@ -325,17 +303,15 @@ class TArrowInferencinator : public NActors::TActorBootstrapped<TArrowInferencin
         auto& arrowFields = std::get<ArrowFields>(mbArrowFields);
         std::vector<Ydb::Column> ydbFields;
         for (const auto& field : arrowFields) {
-            if (field->name().empty()) {
+            Ydb::Column column;
+            if (!ArrowToYdbType(*column.mutable_type(), *field->type(), file.Config)) {
                 continue;
             }
-            ydbFields.emplace_back();
-            auto& ydbField = ydbFields.back();
-            if (!ArrowToYdbType(*ydbField.mutable_type(), *field->type(), file.Config)) {
-                ctx.Send(RequesterId_, MakeErrorSchema(file.Path, NFq::TIssuesIds::UNSUPPORTED, TStringBuilder{} << "couldn't convert arrow type to ydb: " << field->ToString()));
-                RequesterId_ = {};
-                return;
+            if (field->name().empty()) {
+                continue;
             }
-            ydbField.mutable_name()->assign(field->name());
+            column.mutable_name()->assign(field->name());
+            ydbFields.push_back(column);
         }
 
         ctx.Send(RequesterId_, new TEvInferredFileSchema(file.Path, std::move(ydbFields)));

ydb/tests/fq/s3/test_s3_0.py

@@ -624,6 +624,50 @@ def test_inference_null_column_name(self, kikimr, s3, client, unique_prefix):
         assert result_set.rows[2].items[1].int64_value == 15
         assert sum(kikimr.control_plane.get_metering(1)) == 10
 
+    @yq_v2
+    @pytest.mark.parametrize("client", [{"folder_id": "my_folder"}], indirect=True)
+    def test_inference_unsupported_types(self, kikimr, s3, client, unique_prefix):
+        resource = boto3.resource(
+            "s3", endpoint_url=s3.s3_url, aws_access_key_id="key", aws_secret_access_key="secret_key"
+        )
+
+        bucket = resource.Bucket("fbucket")
+        bucket.create(ACL='public-read')
+        bucket.objects.all().delete()
+
+        s3_client = boto3.client(
+            "s3", endpoint_url=s3.s3_url, aws_access_key_id="key", aws_secret_access_key="secret_key"
+        )
+
+        fruits = '''{ "a" : [10, 20, 30] , "b" : { "key" : "value" }, "c" : 10 }
+{ "a" : [10, 20, 30] , "b" : { "key" : "value" }, "c" : 20 }
+{ "a" : [10, 20, 30] , "b" : { "key" : "value" }, "c" : 30 }'''
+        s3_client.put_object(Body=fruits, Bucket='fbucket', Key='fruits.json', ContentType='text/plain')
+        kikimr.control_plane.wait_bootstrap(1)
+        storage_connection_name = unique_prefix + "fruitbucket"
+        client.create_storage_connection(storage_connection_name, "fbucket")
+
+        sql = f'''
+            SELECT *
+            FROM `{storage_connection_name}`.`fruits.json`
+            WITH (format=json_each_row, with_infer='true');
+            '''
+
+        query_id = client.create_query("simple", sql, type=fq.QueryContent.QueryType.ANALYTICS).result.query_id
+        client.wait_query_status(query_id, fq.QueryMeta.COMPLETED)
+
+        data = client.get_result_data(query_id)
+        result_set = data.result.result_set
+        logging.debug(str(result_set))
+        assert len(result_set.columns) == 1
+        assert result_set.columns[0].name == "c"
+        assert result_set.columns[0].type.optional_type.item.type_id == ydb.Type.INT64
+        assert len(result_set.rows) == 3
+        assert result_set.rows[0].items[0].int64_value == 10
+        assert result_set.rows[1].items[0].int64_value == 20
+        assert result_set.rows[2].items[0].int64_value == 30
+        assert sum(kikimr.control_plane.get_metering(1)) == 10
+
     @yq_all
     @pytest.mark.parametrize("client", [{"folder_id": "my_folder"}], indirect=True)
     def test_csv_with_hopping(self, kikimr, s3, client, unique_prefix):