[GitHub] [iceberg] edgarRd commented on a change in pull request #1208: ORC: Add name mapping support

[GitBox <gi...@apache.org>]

edgarRd commented on a change in pull request #1208:
URL: https://github.com/apache/iceberg/pull/1208#discussion_r459805494



##########
File path: orc/src/test/java/org/apache/iceberg/orc/TestORCSchemaUtil.java
##########
@@ -269,4 +278,68 @@ public void testSkipNonIcebergColumns() {
     );
     assertEquals("Schemas must match.", expectedSchema2.asStruct(), icebergSchema2.asStruct());
   }
+
+  @Test
+  public void testHasIds() {
+    Schema schema = new Schema(
+        optional(1, "data", Types.StructType.of(
+            optional(10, "entries", Types.MapType.ofOptional(11, 12, Types.StringType.get(), Types.DateType.get()))
+        )),
+        optional(2, "intCol", Types.IntegerType.get()),
+        optional(3, "longCol", Types.LongType.get()),
+        optional(4, "listCol", Types.ListType.ofOptional(40, Types.DoubleType.get()))
+    );
+
+    TypeDescription orcSchema = ORCSchemaUtil.removeIds(ORCSchemaUtil.convert(schema));
+    assertFalse("Should not have Ids", ORCSchemaUtil.hasIds(orcSchema));
+
+    TypeDescription map2Col = TypeDescription.createMap(TypeDescription.createString(), TypeDescription.createBinary());
+    map2Col.setAttribute(ICEBERG_ID_ATTRIBUTE, "4");
+    orcSchema.addField("map2Col", map2Col);
+    assertTrue("Should have Ids after adding one type with Id", ORCSchemaUtil.hasIds(orcSchema));
+  }
+
+  @Test
+  public void tetsAssignIdsByNameMapping() {
+    Types.StructType structType = Types.StructType.of(
+        required(0, "id", Types.LongType.get()),
+        optional(1, "list_of_maps",
+            Types.ListType.ofOptional(2, Types.MapType.ofOptional(3, 4,
+                Types.StringType.get(),
+                SUPPORTED_PRIMITIVES))),
+        optional(5, "map_of_lists",
+            Types.MapType.ofOptional(6, 7,
+                Types.StringType.get(),
+                Types.ListType.ofOptional(8, SUPPORTED_PRIMITIVES))),
+        required(9, "list_of_lists",
+            Types.ListType.ofOptional(10, Types.ListType.ofOptional(11, SUPPORTED_PRIMITIVES))),
+        required(12, "map_of_maps",
+            Types.MapType.ofOptional(13, 14,
+                Types.StringType.get(),
+                Types.MapType.ofOptional(15, 16,
+                    Types.StringType.get(),
+                    SUPPORTED_PRIMITIVES))),
+        required(17, "list_of_struct_of_nested_types", Types.ListType.ofOptional(19, Types.StructType.of(
+            Types.NestedField.required(20, "m1", Types.MapType.ofOptional(21, 22,
+                Types.StringType.get(),
+                SUPPORTED_PRIMITIVES)),
+            Types.NestedField.optional(23, "l1", Types.ListType.ofRequired(24, SUPPORTED_PRIMITIVES)),
+            Types.NestedField.required(25, "l2", Types.ListType.ofRequired(26, SUPPORTED_PRIMITIVES)),
+            Types.NestedField.optional(27, "m2", Types.MapType.ofOptional(28, 29,
+                Types.StringType.get(),
+                SUPPORTED_PRIMITIVES))
+        )))
+    );
+
+    Schema schema = new Schema(TypeUtil.assignFreshIds(structType, new AtomicInteger(0)::incrementAndGet)
+        .asStructType().fields());
+
+    NameMapping nameMapping = MappingUtil.create(schema);
+    TypeDescription typeDescriptionWithIds = ORCSchemaUtil.convert(schema);
+    TypeDescription typeDescriptionWithIdsFromNameMapping = ORCSchemaUtil
+        .applyNameMapping(ORCSchemaUtil.removeIds(typeDescriptionWithIds), nameMapping);
+
+    assertTrue("TypeDescription schemas should be equal, not comparing Attributes",
+        typeDescriptionWithIds.equals(typeDescriptionWithIdsFromNameMapping, false /* checkAttributes */));

Review comment:
       Name mapping - using an Iceberg schema - will select the columns from an ORC schema that match the columns skipping anything that does not match and adds the ids found to produce the reading schema. So, I think the effect of name mapping is still being tested in this case since we're checking the the `TypeDescription` generated with name mapping is the expected one. That being said, I've added a check for the IDs as well. 




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org