Backport GroupsAccumulator for Duration min/max agg (#288) (apache#15322) v47

Author: svranesevic

datafusion/functions-aggregate/src/min_max.rs

@@ -32,9 +32,10 @@ use arrow::array::{
 };
 use arrow::compute;
 use arrow::datatypes::{
-    DataType, Decimal128Type, Decimal256Type, Float16Type, Float32Type, Float64Type,
-    Int16Type, Int32Type, Int64Type, Int8Type, UInt16Type, UInt32Type, UInt64Type,
-    UInt8Type,
+    DataType, Decimal128Type, Decimal256Type, DurationMicrosecondType,
+    DurationMillisecondType, DurationNanosecondType, DurationSecondType, Float16Type,
+    Float32Type, Float64Type, Int16Type, Int32Type, Int64Type, Int8Type, UInt16Type,
+    UInt32Type, UInt64Type, UInt8Type,
 };
 use arrow_schema::IntervalUnit;
 use datafusion_common::stats::Precision;
@@ -264,6 +265,7 @@ impl AggregateUDFImpl for Max {
                 | Binary
                 | LargeBinary
                 | BinaryView
+                | Duration(_)
         )
     }
 
@@ -318,6 +320,18 @@ impl AggregateUDFImpl for Max {
             Timestamp(Nanosecond, _) => {
                 primitive_max_accumulator!(data_type, i64, TimestampNanosecondType)
             }
+            Duration(Second) => {
+                primitive_max_accumulator!(data_type, i64, DurationSecondType)
+            }
+            Duration(Millisecond) => {
+                primitive_max_accumulator!(data_type, i64, DurationMillisecondType)
+            }
+            Duration(Microsecond) => {
+                primitive_max_accumulator!(data_type, i64, DurationMicrosecondType)
+            }
+            Duration(Nanosecond) => {
+                primitive_max_accumulator!(data_type, i64, DurationNanosecondType)
+            }
             Decimal128(_, _) => {
                 primitive_max_accumulator!(data_type, i128, Decimal128Type)
             }
@@ -1118,6 +1132,7 @@ impl AggregateUDFImpl for Min {
                 | Binary
                 | LargeBinary
                 | BinaryView
+                | Duration(_)
         )
     }
 
@@ -1172,6 +1187,18 @@ impl AggregateUDFImpl for Min {
             Timestamp(Nanosecond, _) => {
                 primitive_min_accumulator!(data_type, i64, TimestampNanosecondType)
             }
+            Duration(Second) => {
+                primitive_min_accumulator!(data_type, i64, DurationSecondType)
+            }
+            Duration(Millisecond) => {
+                primitive_min_accumulator!(data_type, i64, DurationMillisecondType)
+            }
+            Duration(Microsecond) => {
+                primitive_min_accumulator!(data_type, i64, DurationMicrosecondType)
+            }
+            Duration(Nanosecond) => {
+                primitive_min_accumulator!(data_type, i64, DurationNanosecondType)
+            }
             Decimal128(_, _) => {
                 primitive_min_accumulator!(data_type, i128, Decimal128Type)
             }

datafusion/sqllogictest/test_files/aggregate.slt

@@ -4339,8 +4339,8 @@ NULL NULL Row 2 Y
 statement ok
 create table d
 as values
-  (arrow_cast(1, 'Duration(Second)'), arrow_cast(2, 'Duration(Millisecond)'),  arrow_cast(3, 'Duration(Microsecond)'),  arrow_cast(4, 'Duration(Nanosecond)')),
-  (arrow_cast(11, 'Duration(Second)'),arrow_cast(22, 'Duration(Millisecond)'), arrow_cast(33, 'Duration(Microsecond)'), arrow_cast(44, 'Duration(Nanosecond)'));
+  (arrow_cast(1, 'Duration(Second)'), arrow_cast(2, 'Duration(Millisecond)'),  arrow_cast(3, 'Duration(Microsecond)'),  arrow_cast(4, 'Duration(Nanosecond)'), 1),
+  (arrow_cast(11, 'Duration(Second)'),arrow_cast(22, 'Duration(Millisecond)'), arrow_cast(33, 'Duration(Microsecond)'), arrow_cast(44, 'Duration(Nanosecond)'), 1);
 
 query ????
 SELECT min(column1), min(column2), min(column3), min(column4) FROM d;
@@ -4352,6 +4352,32 @@ SELECT max(column1), max(column2), max(column3), max(column4) FROM d;
 ----
 0 days 0 hours 0 mins 11 secs 0 days 0 hours 0 mins 0.022 secs 0 days 0 hours 0 mins 0.000033 secs 0 days 0 hours 0 mins 0.000000044 secs
 
+# GROUP BY follows a different code path
+query ????I
+SELECT min(column1), min(column2), min(column3), min(column4), column5 FROM d GROUP BY column5;
+----
+0 days 0 hours 0 mins 1 secs 0 days 0 hours 0 mins 0.002 secs 0 days 0 hours 0 mins 0.000003 secs 0 days 0 hours 0 mins 0.000000004 secs 1
+
+query ????I
+SELECT max(column1), max(column2), max(column3), max(column4), column5 FROM d GROUP BY column5;
+----
+0 days 0 hours 0 mins 11 secs 0 days 0 hours 0 mins 0.022 secs 0 days 0 hours 0 mins 0.000033 secs 0 days 0 hours 0 mins 0.000000044 secs 1
+
+statement ok
+INSERT INTO d VALUES
+  (arrow_cast(3, 'Duration(Second)'), arrow_cast(1, 'Duration(Millisecond)'), arrow_cast(7, 'Duration(Microsecond)'), arrow_cast(2, 'Duration(Nanosecond)'), 1),
+  (arrow_cast(0, 'Duration(Second)'), arrow_cast(9, 'Duration(Millisecond)'), arrow_cast(5, 'Duration(Microsecond)'), arrow_cast(8, 'Duration(Nanosecond)'), 1);
+
+query ????I
+SELECT max(column1), max(column2), max(column3), max(column4), column5 FROM d GROUP BY column5 ORDER BY column5;
+----
+0 days 0 hours 0 mins 11 secs 0 days 0 hours 0 mins 0.022 secs 0 days 0 hours 0 mins 0.000033 secs 0 days 0 hours 0 mins 0.000000044 secs 1
+
+query ????I
+SELECT min(column1), min(column2), min(column3), min(column4), column5 FROM d GROUP BY column5 ORDER BY column5;
+----
+0 days 0 hours 0 mins 0 secs 0 days 0 hours 0 mins 0.001 secs 0 days 0 hours 0 mins 0.000003 secs 0 days 0 hours 0 mins 0.000000002 secs 1
+
 statement ok
 drop table d;