Make In Memory Table Aggregator types more specific where possible (#3679)

Many aggregation types fell back to the general `Any` type where they could have used the type of input column - for example `First` of a column of integers is guaranteed to fit the `Integer` storage type, so it doesn't have to fall back to `Any`. This PR fixes that and adds a test that checks this.

std-bits/table/src/main/java/org/enso/table/aggregations/First.java

@@ -27,7 +27,7 @@ public class First extends Aggregator {
       Column[] orderByColumns,
       Long[] orderByDirections,
       Comparator<Object> objectComparator) {
-    super(name, Storage.Type.OBJECT);
+    super(name, column.getStorage().getType());
     this.storage = column.getStorage();
     this.orderByColumns =
         orderByColumns == null

std-bits/table/src/main/java/org/enso/table/aggregations/GroupBy.java

@@ -10,7 +10,7 @@ public class GroupBy extends Aggregator {
   private final Storage storage;
 
   public GroupBy(String name, Column column) {
-    super(name, Storage.Type.OBJECT);
+    super(name, column.getStorage().getType());
     storage = column.getStorage();
   }
 

std-bits/table/src/main/java/org/enso/table/aggregations/Last.java

@@ -26,7 +26,7 @@ public class Last extends Aggregator {
       Column[] orderByColumns,
       Long[] orderByDirections,
       Comparator<Object> objectComparator) {
-    super(name, Storage.Type.OBJECT);
+    super(name, column.getStorage().getType());
     this.storage = column.getStorage();
     this.orderByColumns =
         orderByColumns == null

std-bits/table/src/main/java/org/enso/table/aggregations/MinOrMax.java

@@ -23,7 +23,7 @@ public class MinOrMax extends Aggregator {
    * @param minOrMax <0 for minimum, >0 for maximum
    */
   public MinOrMax(String name, Column column, int minOrMax, Comparator<Object> objectComparator) {
-    super(name, Storage.Type.OBJECT);
+    super(name, column.getStorage().getType());
     this.storage = column.getStorage();
     this.minOrMax = Integer.signum(minOrMax);
     this.objectComparator = objectComparator;

std-bits/table/src/main/java/org/enso/table/aggregations/Mode.java

@@ -13,7 +13,7 @@ public class Mode extends Aggregator {
   private final Storage storage;
 
   public Mode(String name, Column column) {
-    super(name, Storage.Type.OBJECT);
+    super(name, column.getStorage().getType());
     this.storage = column.getStorage();
   }
 

std-bits/table/src/main/java/org/enso/table/data/column/storage/BoolStorage.java

@@ -39,7 +39,7 @@ public class BoolStorage extends Storage {
   }
 
   @Override
-  public long getType() {
+  public int getType() {
     return Type.BOOL;
   }
 

std-bits/table/src/main/java/org/enso/table/data/column/storage/DateStorage.java

@@ -30,7 +30,7 @@ public class DateStorage extends SpecializedStorage<LocalDate> {
   }
 
   @Override
-  public long getType() {
+  public int getType() {
     return Type.DATE;
   }
 }

std-bits/table/src/main/java/org/enso/table/data/column/storage/DateTimeStorage.java

@@ -30,7 +30,7 @@ public class DateTimeStorage extends SpecializedStorage<ZonedDateTime> {
   }
 
   @Override
-  public long getType() {
+  public int getType() {
     return Type.DATE_TIME;
   }
 }

std-bits/table/src/main/java/org/enso/table/data/column/storage/DoubleStorage.java

@@ -63,7 +63,7 @@ public class DoubleStorage extends NumericStorage {
 
   /** @inheritDoc */
   @Override
-  public long getType() {
+  public int getType() {
     return Type.DOUBLE;
   }
 

std-bits/table/src/main/java/org/enso/table/data/column/storage/LongStorage.java

@@ -71,7 +71,7 @@ public class LongStorage extends NumericStorage {
 
   /** @inheritDoc */
   @Override
-  public long getType() {
+  public int getType() {
     return Type.LONG;
   }
 

std-bits/table/src/main/java/org/enso/table/data/column/storage/ObjectStorage.java

@@ -26,7 +26,7 @@ public class ObjectStorage extends SpecializedStorage<Object> {
   }
 
   @Override
-  public long getType() {
+  public int getType() {
     return Type.OBJECT;
   }
 

std-bits/table/src/main/java/org/enso/table/data/column/storage/SpecializedStorage.java

@@ -15,7 +15,7 @@ public abstract class SpecializedStorage<T> extends Storage {
   protected abstract T[] newUnderlyingArray(int size);
 
   @Override
-  public abstract long getType();
+  public abstract int getType();
 
   /**
    * @param data the underlying data

std-bits/table/src/main/java/org/enso/table/data/column/storage/Storage.java

@@ -24,7 +24,7 @@ public abstract class Storage {
   public abstract int countMissing();
 
   /** @return the type tag of this column's storage. Must be one of {@link Type} */
-  public abstract long getType();
+  public abstract int getType();
 
   /**
    * Checks whether the value at {@code idx} is missing.

std-bits/table/src/main/java/org/enso/table/data/column/storage/StringStorage.java

@@ -29,7 +29,7 @@ public class StringStorage extends SpecializedStorage<String> {
   }
 
   @Override
-  public long getType() {
+  public int getType() {
     return Type.STRING;
   }
 

std-bits/table/src/main/java/org/enso/table/data/column/storage/TimeOfDayStorage.java

@@ -30,7 +30,7 @@ public class TimeOfDayStorage extends SpecializedStorage<LocalTime> {
   }
 
   @Override
-  public long getType() {
+  public int getType() {
     return Type.TIME_OF_DAY;
   }
 }

std-bits/table/src/main/java/org/enso/table/data/index/MultiValueIndex.java

@@ -102,11 +102,14 @@ public class MultiValueIndex {
 
   private static Builder getBuilderForType(int type, int size) {
     return switch (type) {
-      case Storage.Type.BOOL -> new BoolBuilder();
-      case Storage.Type.DOUBLE -> NumericBuilder.createDoubleBuilder(size);
-      case Storage.Type.LONG -> NumericBuilder.createLongBuilder(size);
-      case Storage.Type.STRING -> new StringBuilder(size);
       case Storage.Type.OBJECT -> new ObjectBuilder(size);
+      case Storage.Type.LONG -> NumericBuilder.createLongBuilder(size);
+      case Storage.Type.DOUBLE -> NumericBuilder.createDoubleBuilder(size);
+      case Storage.Type.STRING -> new StringBuilder(size);
+      case Storage.Type.BOOL -> new BoolBuilder();
+      case Storage.Type.DATE -> new DateBuilder(size);
+      case Storage.Type.TIME_OF_DAY -> new TimeOfDayBuilder(size);
+      case Storage.Type.DATE_TIME -> new DateTimeBuilder(size);
       default -> new InferredBuilder(size);
     };
   }

test/Table_Tests/src/Table_Spec.enso

@@ -5,6 +5,8 @@ from Standard.Table import Column, Sort_Column, Sort_Column_Selector
 from Standard.Table.Data.Table import Empty_Error
 from Standard.Table.Errors as Table_Errors import Invalid_Output_Column_Names_Data, Duplicate_Output_Column_Names_Data, No_Input_Columns_Selected, Missing_Input_Columns_Data
 import Standard.Table.Data.Storage
+import Standard.Table.Data.Aggregate_Column
+from Standard.Table.Data.Aggregate_Column import all hiding First, Last
 
 import Standard.Visualization
 
@@ -617,4 +619,31 @@ spec =
             problems = [Duplicate_Output_Column_Names_Data ["A", "A", "A"]]
             Problems.test_problem_handling action problems tester
 
+    Test.group "[In-Memory] Table.aggregate" <|
+        Test.specify "should return columns with correct types" <|
+            dates = ["dates", [Date.new 1999, Date.new 2000, Date.new 2000, Date.new 2000]]
+            texts = ["texts", ["a", "bb", "a", "bb"]]
+            mixed = ["mixed", [1, "a", "a", 1]]
+            ints = ["ints", [0, 1, 1, 0]]
+            floats = ["floats", [0.1, 1.0, 2.0, 1.5]]
+            objects = ["objects", [My_Data 0 1, My_Data 0 1, My_Data 2 2, My_Data 2 2]]
+            table = Table.new [dates, texts, mixed, ints, floats, objects]
+
+            t1 = table.aggregate [Group_By "dates", Shortest "texts", Aggregate_Column.First "texts", Aggregate_Column.First "objects", Aggregate_Column.First "ints", Aggregate_Column.Last "mixed"]
+            t1.info.at "Column" . to_vector . should_equal ["dates", "Shortest texts", "First texts", "First objects", "First ints", "Last mixed"]
+            t1.info.at "Storage Type" . to_vector . should_equal [Storage.Date, Storage.Text, Storage.Text, Storage.Any, Storage.Integer, Storage.Any]
+
+            t2 = table.aggregate [Mode "dates", Count_Distinct "objects", Count_Distinct "texts", Minimum "ints", Maximum "floats"]
+            t2.info.at "Column" . to_vector . should_equal ["Mode dates", "Count Distinct objects", "Count Distinct texts", "Minimum ints", "Maximum floats"]
+            t2.info.at "Storage Type" . to_vector . should_equal [Storage.Date, Storage.Integer, Storage.Integer, Storage.Integer, Storage.Decimal]
+
+            t3 = table.aggregate [Group_By "texts", Group_By "ints", Aggregate_Column.Last "floats"]
+            t3.info.at "Column" . to_vector . should_equal ["texts", "ints", "Last floats"]
+            t3.info.at "Storage Type" . to_vector . should_equal [Storage.Text, Storage.Integer, Storage.Decimal]
+
+            t4 = table.aggregate [Group_By "mixed", Sum "ints", Sum "floats"]
+            t4.info.at "Column" . to_vector . should_equal ["mixed", "Sum ints", "Sum floats"]
+            t4.info.at "Storage Type" . to_vector . should_equal [Storage.Any, Storage.Decimal, Storage.Decimal]
+
+
 main = Test.Suite.run_main spec