Github user marmbrus commented on a diff in the pull request:
https://github.com/apache/spark/pull/9567#discussion_r44323839
--- Diff:
sql/core/src/test/java/test/org/apache/spark/sql/JavaDatasetSuite.java ---
@@ -299,59 +299,33 @@ public void testJoin() {
@Test
public void testTupleEncoder() {
- Encoder<Tuple2<Integer, String>> encoder2 = e.tuple(e.INT(),
e.STRING());
+ Encoder<Tuple2<Integer, String>> encoder2 = e.tuple(Integer.class,
String.class);
List<Tuple2<Integer, String>> data2 = Arrays.asList(tuple2(1, "a"),
tuple2(2, "b"));
Dataset<Tuple2<Integer, String>> ds2 = context.createDataset(data2,
encoder2);
Assert.assertEquals(data2, ds2.collectAsList());
- Encoder<Tuple3<Integer, Long, String>> encoder3 = e.tuple(e.INT(),
e.LONG(), e.STRING());
+ Encoder<Tuple3<Integer, Long, String>> encoder3 =
+ e.tuple(Integer.class, Long.class, String.class);
List<Tuple3<Integer, Long, String>> data3 =
Arrays.asList(new Tuple3<Integer, Long, String>(1, 2L, "a"));
Dataset<Tuple3<Integer, Long, String>> ds3 =
context.createDataset(data3, encoder3);
Assert.assertEquals(data3, ds3.collectAsList());
Encoder<Tuple4<Integer, String, Long, String>> encoder4 =
- e.tuple(e.INT(), e.STRING(), e.LONG(), e.STRING());
+ e.tuple(Integer.class, String.class, Long.class, String.class);
List<Tuple4<Integer, String, Long, String>> data4 =
Arrays.asList(new Tuple4<Integer, String, Long, String>(1, "b", 2L,
"a"));
Dataset<Tuple4<Integer, String, Long, String>> ds4 =
context.createDataset(data4, encoder4);
Assert.assertEquals(data4, ds4.collectAsList());
Encoder<Tuple5<Integer, String, Long, String, Boolean>> encoder5 =
- e.tuple(e.INT(), e.STRING(), e.LONG(), e.STRING(), e.BOOLEAN());
+ e.tuple(Integer.class, String.class, Long.class, String.class,
Boolean.class);
List<Tuple5<Integer, String, Long, String, Boolean>> data5 =
Arrays.asList(new Tuple5<Integer, String, Long, String, Boolean>(1,
"b", 2L, "a", true));
Dataset<Tuple5<Integer, String, Long, String, Boolean>> ds5 =
context.createDataset(data5, encoder5);
Assert.assertEquals(data5, ds5.collectAsList());
- }
- @Test
- public void testNestedTupleEncoder() {
- // test ((int, string), string)
- Encoder<Tuple2<Tuple2<Integer, String>, String>> encoder =
- e.tuple(e.tuple(e.INT(), e.STRING()), e.STRING());
- List<Tuple2<Tuple2<Integer, String>, String>> data =
- Arrays.asList(tuple2(tuple2(1, "a"), "a"), tuple2(tuple2(2, "b"),
"b"));
- Dataset<Tuple2<Tuple2<Integer, String>, String>> ds =
context.createDataset(data, encoder);
- Assert.assertEquals(data, ds.collectAsList());
-
- // test (int, (string, string, long))
- Encoder<Tuple2<Integer, Tuple3<String, String, Long>>> encoder2 =
- e.tuple(e.INT(), e.tuple(e.STRING(), e.STRING(), e.LONG()));
- List<Tuple2<Integer, Tuple3<String, String, Long>>> data2 =
- Arrays.asList(tuple2(1, new Tuple3<String, String, Long>("a", "b",
3L)));
- Dataset<Tuple2<Integer, Tuple3<String, String, Long>>> ds2 =
- context.createDataset(data2, encoder2);
- Assert.assertEquals(data2, ds2.collectAsList());
-
- // test (int, ((string, long), string))
- Encoder<Tuple2<Integer, Tuple2<Tuple2<String, Long>, String>>>
encoder3 =
- e.tuple(e.INT(), e.tuple(e.tuple(e.STRING(), e.LONG()), e.STRING()));
- List<Tuple2<Integer, Tuple2<Tuple2<String, Long>, String>>> data3 =
- Arrays.asList(tuple2(1, tuple2(tuple2("a", 2L), "b")));
- Dataset<Tuple2<Integer, Tuple2<Tuple2<String, Long>, String>>> ds3 =
- context.createDataset(data3, encoder3);
- Assert.assertEquals(data3, ds3.collectAsList());
+ // TODO: support nested tuple encoder in java.
--- End diff --
It seems this is very hard unless we expose the API for creating Tuples
from encoders?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
