andygrove commented on code in PR #2750:
URL: https://github.com/apache/arrow-datafusion/pull/2750#discussion_r903138887
##########
datafusion/core/tests/sql/joins.rs:
##########
@@ -1342,3 +1342,458 @@ async fn join_with_hash_unsupported_data_type() ->
Result<()> {
Ok(())
}
+
+#[tokio::test]
+async fn reduce_left_join() -> Result<()> {
+ let ctx = SessionContext::new();
+
+ let t1_schema = Schema::new(vec![
+ Field::new("c1", DataType::Int32, true),
+ Field::new("c2", DataType::Utf8, true),
+ Field::new("c3", DataType::Int64, true),
+ ]);
+ let t1_data = RecordBatch::try_new(
+ Arc::new(t1_schema),
+ vec![
+ Arc::new(Int32Array::from_slice(&[1, 2, 3])),
+ Arc::new(StringArray::from_slice(&["aaa", "bbb", "ccc"])),
+ Arc::new(Int64Array::from_slice(&[100, 200, 300])),
+ ],
+ )?;
+ let t1 = MemTable::try_new(t1_data.schema(), vec![vec![t1_data]])?;
+ ctx.register_table("t1", Arc::new(t1))?;
+
+ let t2_schema = Schema::new(vec![
+ Field::new("c4", DataType::Int32, true),
+ Field::new("c5", DataType::Utf8, true),
+ Field::new("c6", DataType::Int64, true),
+ ]);
+ let t2_data = RecordBatch::try_new(
+ Arc::new(t2_schema),
+ vec![
+ Arc::new(Int32Array::from_slice(&[3, 4, 5])),
+ Arc::new(StringArray::from_slice(&["ccc", "ddd", "eee"])),
+ Arc::new(Int64Array::from_slice(&[300, 400, 500])),
+ ],
+ )?;
+ let t2 = MemTable::try_new(t2_data.schema(), vec![vec![t2_data]])?;
+ ctx.register_table("t2", Arc::new(t2))?;
+
+ // reduce to inner join
+ let sql = "select * from t1 left join t2 on t1.c1 = t2.c4 where t2.c6 <
1000";
+ let msg = format!("Creating logical plan for '{}'", sql);
+ let plan = ctx
+ .create_logical_plan(&("explain ".to_owned() + sql))
+ .expect(&msg);
+ let state = ctx.state();
+ let plan = state.optimize(&plan)?;
+ let expected = vec![
+ "Explain [plan_type:Utf8, plan:Utf8]",
+ " Projection: #t1.c1, #t1.c2, #t1.c3, #t2.c4, #t2.c5, #t2.c6
[c1:Int32;N, c2:Utf8;N, c3:Int64;N, c4:Int32;N, c5:Utf8;N, c6:Int64;N]",
+ " Inner Join: #t1.c1 = #t2.c4 [c1:Int32;N, c2:Utf8;N, c3:Int64;N,
c4:Int32;N, c5:Utf8;N, c6:Int64;N]",
+ " TableScan: t1 projection=Some([c1, c2, c3]) [c1:Int32;N,
c2:Utf8;N, c3:Int64;N]",
+ " Filter: #t2.c6 < Int64(1000) [c4:Int32;N, c5:Utf8;N,
c6:Int64;N]",
+ " TableScan: t2 projection=Some([c4, c5, c6]) [c4:Int32;N,
c5:Utf8;N, c6:Int64;N]",
+ ];
+ let formatted = plan.display_indent_schema().to_string();
+ let actual: Vec<&str> = formatted.trim().lines().collect();
+ assert_eq!(
+ expected, actual,
+ "\n\nexpected:\n\n{:#?}\nactual:\n\n{:#?}\n\n",
+ expected, actual
+ );
+ let expected = vec![
+ "+----+-----+-----+----+-----+-----+",
+ "| c1 | c2 | c3 | c4 | c5 | c6 |",
+ "+----+-----+-----+----+-----+-----+",
+ "| 3 | ccc | 300 | 3 | ccc | 300 |",
+ "+----+-----+-----+----+-----+-----+",
+ ];
+
+ let results = execute_to_batches(&ctx, sql).await;
+ assert_batches_sorted_eq!(expected, &results);
+
+ // could not reduce, use left join
Review Comment:
Can we split the tests out into separate test methods?
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
