fix: distinct respect in range (#5015)

* fix: distinct respect in range * tests: sqlness * chore: newline
GreptimeTeam · Nov 18, 2024 · 4402f63 · 4402f63
1 parent c199604
commit 4402f63
Show file tree

Hide file tree

Showing 3 changed files with 80 additions and 2 deletions.
diff --git a/src/query/src/range_select/plan.rs b/src/query/src/range_select/plan.rs
@@ -773,6 +773,8 @@ impl RangeSelect {
                         } else {
                             vec![]
                         };
+                        let distinct = aggr.distinct;
+                        // TODO(discord9): add default null treatment?
 
                         let input_phy_exprs = self.create_physical_expr_list(
                             matches!(
@@ -786,7 +788,7 @@ impl RangeSelect {
                         match &aggr.func_def {
                             AggregateFunctionDefinition::BuiltIn(fun) => create_aggr_expr(
                                 fun,
-                                false,
+                                distinct,
                                 &input_phy_exprs,
                                 &order_by,
                                 &input_schema,
@@ -801,7 +803,7 @@ impl RangeSelect {
                                 &input_schema,
                                 name,
                                 false,
-                                false,
+                                distinct,
                             ),
                         }
                     }

diff --git a/tests/cases/standalone/common/range/by.result b/tests/cases/standalone/common/range/by.result
@@ -105,3 +105,53 @@ DROP TABLE host;
 
 Affected Rows: 0
 
+CREATE TABLE grpc_latencies (
+  ts TIMESTAMP TIME INDEX,
+  host STRING,
+  method_name STRING,
+  latency DOUBLE,
+  PRIMARY KEY (host, method_name)
+) with('append_mode'='true');
+
+Affected Rows: 0
+
+INSERT INTO grpc_latencies (ts, host, method_name, latency) VALUES
+  ('2024-07-11 20:00:06', 'host1', 'GetUser', 103.0),
+  ('2024-07-11 20:00:06', 'host2', 'GetUser', 113.0),
+  ('2024-07-11 20:00:07', 'host1', 'GetUser', 103.5),
+  ('2024-07-11 20:00:07', 'host2', 'GetUser', 107.0),
+  ('2024-07-11 20:00:08', 'host1', 'GetUser', 104.0),
+  ('2024-07-11 20:00:08', 'host2', 'GetUser', 96.0),
+  ('2024-07-11 20:00:09', 'host1', 'GetUser', 104.5),
+  ('2024-07-11 20:00:09', 'host2', 'GetUser', 114.0);
+
+Affected Rows: 8
+
+SELECT ts, count(distinct host) RANGE '5s' as h FROM grpc_latencies ALIGN '5s' by (method_name);
+
++---------------------+---+
+| ts                  | h |
++---------------------+---+
+| 2024-07-11T20:00:05 | 2 |
++---------------------+---+
+
+SELECT ts, count(*) RANGE '5s' as h FROM grpc_latencies ALIGN '5s' by (method_name);
+
++---------------------+---+
+| ts                  | h |
++---------------------+---+
+| 2024-07-11T20:00:05 | 8 |
++---------------------+---+
+
+select date_bin(INTERVAL '5s', ts) as t, count(distinct host) as count from grpc_latencies group by t;
+
++---------------------+-------+
+| t                   | count |
++---------------------+-------+
+| 2024-07-11T20:00:05 | 2     |
++---------------------+-------+
+
+DROP TABLE grpc_latencies;
+
+Affected Rows: 0
+
diff --git a/tests/cases/standalone/common/range/by.sql b/tests/cases/standalone/common/range/by.sql
@@ -58,3 +58,29 @@ INSERT INTO TABLE host VALUES
 SELECT ts, max(val) RANGE '5s' FROM host ALIGN '20s' ORDER BY ts;
 
 DROP TABLE host;
+
+CREATE TABLE grpc_latencies (
+  ts TIMESTAMP TIME INDEX,
+  host STRING,
+  method_name STRING,
+  latency DOUBLE,
+  PRIMARY KEY (host, method_name)
+) with('append_mode'='true');
+
+INSERT INTO grpc_latencies (ts, host, method_name, latency) VALUES
+  ('2024-07-11 20:00:06', 'host1', 'GetUser', 103.0),
+  ('2024-07-11 20:00:06', 'host2', 'GetUser', 113.0),
+  ('2024-07-11 20:00:07', 'host1', 'GetUser', 103.5),
+  ('2024-07-11 20:00:07', 'host2', 'GetUser', 107.0),
+  ('2024-07-11 20:00:08', 'host1', 'GetUser', 104.0),
+  ('2024-07-11 20:00:08', 'host2', 'GetUser', 96.0),
+  ('2024-07-11 20:00:09', 'host1', 'GetUser', 104.5),
+  ('2024-07-11 20:00:09', 'host2', 'GetUser', 114.0);
+
+SELECT ts, count(distinct host) RANGE '5s' as h FROM grpc_latencies ALIGN '5s' by (method_name);
+
+SELECT ts, count(*) RANGE '5s' as h FROM grpc_latencies ALIGN '5s' by (method_name);
+
+select date_bin(INTERVAL '5s', ts) as t, count(distinct host) as count from grpc_latencies group by t;
+
+DROP TABLE grpc_latencies;