Add Custom SQL to In Database aggregation (#14472)

- Allows user to choose Custom SQL in an aggregate.

https://github.com/user-attachments/assets/badee3cf-c35c-430d-9c25-3d26d15c316c

Author: jdunkerley

CHANGELOG.md

@@ -93,6 +93,7 @@
 - [Add Text_Column.index_of][14428]
 - [Trigonometry and other maths function on Column.][14433]
 - [Support for reading JSON lines files.][14439]
+- [Add Custom SQL to in database aggreates.][14472]
 
 [13769]: https://github.com/enso-org/enso/pull/13769
 [14026]: https://github.com/enso-org/enso/pull/14026
@@ -123,6 +124,7 @@
 [14428]: https://github.com/enso-org/enso/pull/14428
 [14433]: https://github.com/enso-org/enso/pull/14433
 [14439]: https://github.com/enso-org/enso/pull/14439
+[14472]: https://github.com/enso-org/enso/pull/14472
 
 #### Enso Language & Runtime
 

distribution/lib/Standard/Database/0.0.0-dev/docs/api/Errors.md

@@ -1,6 +1,6 @@
 ## Enso Signatures 1.0
 ## module Standard.Database.Errors
-- type Aggregagtion_Requires_Order
+- type Aggregation_Requires_Order
     - Error op_name:Standard.Base.Data.Text.Text
     - to_display_text self -> Standard.Base.Data.Text.Text
     - to_text self -> Standard.Base.Data.Text.Text

distribution/lib/Standard/Database/0.0.0-dev/src/Errors.enso

@@ -68,7 +68,7 @@ type Different_Connections
     to_display_text self -> Text =
         self.to_text
 
-type Aggregagtion_Requires_Order
+type Aggregation_Requires_Order
     ## ---
        private: true
        ---

distribution/lib/Standard/Database/0.0.0-dev/src/Internal/Aggregate_Helper.enso

@@ -7,7 +7,6 @@ import Standard.Table.Internal.Aggregate_Column_Helper
 import Standard.Table.Internal.Aggregate_Column_Helper.Internal_Order_By_Column_Reference
 import Standard.Table.Internal.Problem_Builder.Problem_Builder
 from Standard.Table import Aggregate_Column, Table
-from Standard.Table.Aggregate_Column.Aggregate_Column import all
 from Standard.Table.Errors import No_Output_Columns
 
 import project.DB_Table.DB_Table
@@ -18,7 +17,7 @@ import project.Internal.IR.Internal_Column.Internal_Column
 import project.Internal.IR.SQL_IR_Expression.SQL_IR_Expression
 import project.Internal.IR.SQL_IR_Source.SQL_IR_Source
 import project.Internal.SQL_Type_Reference.SQL_Type_Reference
-from project.Errors import Aggregagtion_Requires_Order
+from project.Errors import Aggregation_Requires_Order
 
 ## ---
    private: true
@@ -42,6 +41,12 @@ make_aggregate_column table aggregate as dialect infer_return_type problem_build
         sql_type_ref = infer_return_type op_kind columns expression
         Internal_Column.Value as sql_type_ref expression
 
+    custom_sql function columns =
+        expression = dialect.cast_aggregate_columns "CUSTOM_SQL" columns
+        with_function = SQL_IR_Expression.Operation "CUSTOM_SQL" [SQL_IR_Expression.Literal function]+expression.expressions expression.metadata
+        sql_type_ref = infer_return_type "CUSTOM_SQL" columns with_function
+        Internal_Column.Value as sql_type_ref with_function
+
     aggregate_with_order_by op_kind column order_by =
         order_bys = order_by.map sc->
             effective_ordering = if sc.column.value_type.is_text then Text_Ordering.Default else Nothing
@@ -108,6 +113,7 @@ make_aggregate_column table aggregate as dialect infer_return_type problem_build
         Aggregate_Column.Sum c _ -> simple_aggregate "SUM" [c]
         Aggregate_Column.Average c _ -> simple_aggregate "AVG" [c]
         Aggregate_Column.Median c _ -> simple_aggregate "MEDIAN" [c]
+        Aggregate_Column.SQL fn columns _ -> custom_sql fn columns
 
 ## ---
    private: true
@@ -266,27 +272,27 @@ map_column_inputs f:Function aggregate_column:Aggregate_Column -> Aggregate_Colu
             Internal_Order_By_Column_Reference.Value c direction -> Internal_Order_By_Column_Reference.Value (f c) direction
 
     case aggregate_column of
-        Group_By c as -> Group_By (f c) as
-        Count as -> Count as
-        Count_Distinct c as ignore_nothing ->
-            Count_Distinct ((c:Vector).map f) as ignore_nothing
-        Count_Not_Nothing c as -> Count_Not_Nothing (f c) as
-        Count_Nothing c as -> Count_Nothing (f c) as
-        Count_Not_Empty c as -> Count_Not_Empty (f c) as
-        Count_Empty c as ->  Count_Empty (f c) as
-        Sum c as -> Sum (f c) as
-        Average c as -> Average (f c) as
-        Median c as -> Median (f c) as
-        Percentile p c as -> Percentile p (f c) as
-        Mode c as -> Mode (f c) as
-        Standard_Deviation c as population -> Standard_Deviation (f c) as population
-        Concatenate c as separator prefix suffix quote_char -> Concatenate (f c) as separator prefix suffix quote_char
-        First c as ignore_nothing order_by -> First (f c) as ignore_nothing (update_order_by order_by)
-        Last c as ignore_nothing order_by -> Last (f c) as ignore_nothing (update_order_by order_by)
-        Maximum c as -> Maximum (f c) as
-        Minimum c as -> Minimum (f c) as
-        Shortest c as -> Shortest (f c) as
-        Longest c as -> Longest (f c) as
+        Aggregate_Column.Group_By c as -> Aggregate_Column.Group_By (f c) as
+        Aggregate_Column.Count as -> Aggregate_Column.Count as
+        Aggregate_Column.Count_Distinct c as ignore_nothing ->
+            Aggregate_Column.Count_Distinct ((c:Vector).map f) as ignore_nothing
+        Aggregate_Column.Count_Not_Nothing c as -> Aggregate_Column.Count_Not_Nothing (f c) as
+        Aggregate_Column.Count_Nothing c as -> Aggregate_Column.Count_Nothing (f c) as
+        Aggregate_Column.Count_Not_Empty c as -> Aggregate_Column.Count_Not_Empty (f c) as
+        Aggregate_Column.Count_Empty c as -> Aggregate_Column.Count_Empty (f c) as
+        Aggregate_Column.Sum c as -> Aggregate_Column.Sum (f c) as
+        Aggregate_Column.Average c as -> Aggregate_Column.Average (f c) as
+        Aggregate_Column.Median c as -> Aggregate_Column.Median (f c) as
+        Aggregate_Column.Percentile p c as -> Aggregate_Column.Percentile p (f c) as
+        Aggregate_Column.Mode c as -> Aggregate_Column.Mode (f c) as
+        Aggregate_Column.Standard_Deviation c as population -> Aggregate_Column.Standard_Deviation (f c) as population
+        Aggregate_Column.Concatenate c as separator prefix suffix quote_char -> Aggregate_Column.Concatenate (f c) as separator prefix suffix quote_char
+        Aggregate_Column.First c as ignore_nothing order_by -> Aggregate_Column.First (f c) as ignore_nothing (update_order_by order_by)
+        Aggregate_Column.Last c as ignore_nothing order_by -> Aggregate_Column.Last (f c) as ignore_nothing (update_order_by order_by)
+        Aggregate_Column.Maximum c as -> Aggregate_Column.Maximum (f c) as
+        Aggregate_Column.Minimum c as -> Aggregate_Column.Minimum (f c) as
+        Aggregate_Column.Shortest c as -> Aggregate_Column.Shortest (f c) as
+        Aggregate_Column.Longest c as -> Aggregate_Column.Longest (f c) as
 
 ## ---
    private: true
@@ -296,4 +302,4 @@ is_non_empty_selector v = v.is_nothing.not && v.not_empty
 ## ---
    private: true
    ---
-throw_ordering_required op_name = Error.throw (Aggregagtion_Requires_Order.Error op_name)
+throw_ordering_required op_name = Error.throw (Aggregation_Requires_Order.Error op_name)

distribution/lib/Standard/Database/0.0.0-dev/src/Internal/Postgres/Postgres_Dialect.enso

@@ -8,7 +8,6 @@ import Standard.Base.Errors.Unimplemented.Unimplemented
 import Standard.Table.In_Memory_Column.In_Memory_Column
 import Standard.Table.Internal.Problem_Builder.Problem_Builder
 from Standard.Table import Aggregate_Column, Column, Table, Value_Type
-from Standard.Table.Aggregate_Column.Aggregate_Column import all
 from Standard.Table.Errors import Inexact_Type_Coercion
 
 import project.DB_Column.DB_Column

distribution/lib/Standard/Database/0.0.0-dev/src/Internal/SQLite/SQLite_Dialect.enso

@@ -7,7 +7,6 @@ import Standard.Base.Runtime.Ref.Ref
 
 import Standard.Table.Internal.Problem_Builder.Problem_Builder
 from Standard.Table import Aggregate_Column, Table, Value_Type
-from Standard.Table.Aggregate_Column.Aggregate_Column import all
 
 import project.DB_Column.DB_Column
 import project.DB_Table.DB_Table
@@ -275,34 +274,34 @@ type SQLite_Dialect
        ---
     check_aggregate_support : Aggregate_Column -> Boolean ! Unsupported_Database_Operation
     check_aggregate_support self aggregate = case aggregate of
-        Group_By _ _ -> True
-        Count _ -> True
-        Count_Distinct columns _ _ ->
+        Aggregate_Column.Group_By _ _ -> True
+        Aggregate_Column.Count _ -> True
+        Aggregate_Column.Count_Distinct columns _ _ ->
             if columns.length == 1 then True else
                 unsupported "Count_Distinct on multiple columns"
-        Count_Not_Nothing _ _ -> True
-        Count_Nothing _ _ -> True
-        Count_Not_Empty _ _ -> True
-        Count_Empty _ _ -> True
-        Percentile _ _ _ -> unsupported "Percentile"
-        Mode _ _ -> unsupported "Mode"
-        First _ _ ignore_nothing order_by ->
+        Aggregate_Column.Count_Not_Nothing _ _ -> True
+        Aggregate_Column.Count_Nothing _ _ -> True
+        Aggregate_Column.Count_Not_Empty _ _ -> True
+        Aggregate_Column.Count_Empty _ _ -> True
+        Aggregate_Column.Percentile _ _ _ -> unsupported "Percentile"
+        Aggregate_Column.Mode _ _ -> unsupported "Mode"
+        Aggregate_Column.First _ _ ignore_nothing order_by ->
             if ignore_nothing then unsupported "First with ignore_nothing=True" else
                 if Aggregate_Helper.is_non_empty_selector order_by . not then Aggregate_Helper.throw_ordering_required "First" else
                     True
-        Last _ _ ignore_nothing order_by ->
+        Aggregate_Column.Last _ _ ignore_nothing order_by ->
             if ignore_nothing then unsupported "Last with ignore_nothing=True" else
                 if Aggregate_Helper.is_non_empty_selector order_by . not then Aggregate_Helper.throw_ordering_required "Last" else
                     True
-        Maximum _ _ -> True
-        Minimum _ _ -> True
-        Shortest _ _ -> unsupported "Shortest"
-        Longest _ _ -> unsupported "Longest"
-        Standard_Deviation _ _ _ -> True
-        Concatenate _ _ _ _ _ _ -> True
-        Sum _ _ -> True
-        Average _ _ -> True
-        Median _ _ -> unsupported "Median"
+        Aggregate_Column.Maximum _ _ -> True
+        Aggregate_Column.Minimum _ _ -> True
+        Aggregate_Column.Shortest _ _ -> unsupported "Shortest"
+        Aggregate_Column.Longest _ _ -> unsupported "Longest"
+        Aggregate_Column.Standard_Deviation _ _ _ -> True
+        Aggregate_Column.Concatenate _ _ _ _ _ _ -> True
+        Aggregate_Column.Sum _ _ -> True
+        Aggregate_Column.Average _ _ -> True
+        Aggregate_Column.Median _ _ -> unsupported "Median"
 
     ## ---
        private: true

distribution/lib/Standard/Microsoft/0.0.0-dev/src/Internal/SQLServer_Dialect.enso

@@ -7,7 +7,6 @@ import Standard.Base.Errors.Unimplemented.Unimplemented
 
 import Standard.Table.Internal.Problem_Builder.Problem_Builder
 from Standard.Table import Aggregate_Column, Column, Table, Value_Type
-from Standard.Table.Aggregate_Column.Aggregate_Column import all
 from Standard.Table.Errors import Inexact_Type_Coercion
 
 import Standard.Database.Connection.Connection.Connection
@@ -273,28 +272,28 @@ type SQLServer_Dialect
         unsupported name =
             Error.throw (Unsupported_Database_Operation.Error name) 
         case aggregate of
-            Group_By _ _ -> True
-            Count _ -> True
-            Count_Distinct columns _ _ ->
+            Aggregate_Column.Group_By _ _ -> True
+            Aggregate_Column.Count _ -> True
+            Aggregate_Column.Count_Distinct columns _ _ ->
                 if columns.length == 1 then True else
                     unsupported "Count_Distinct on multiple columns"
-            Count_Not_Nothing _ _ -> True
-            Count_Nothing _ _ -> True
-            Count_Not_Empty _ _ -> True
-            Count_Empty _ _ -> True
-            Percentile _ _ _ -> unsupported "Percentile"
-            Mode _ _ -> unsupported "Mode"
-            First _ _ _ _ -> unsupported "First"
-            Last _ _ _ _ -> unsupported "Last"
-            Maximum _ _ -> True
-            Minimum _ _ -> True
-            Shortest _ _ -> unsupported "Shortest"
-            Longest _ _ -> unsupported "Longest"
-            Standard_Deviation _ _ _ -> True
-            Concatenate _ _ _ _ _ _ -> True
-            Sum _ _ -> True
-            Average _ _ -> True
-            Median _ _ -> unsupported "Median"
+            Aggregate_Column.Count_Not_Nothing _ _ -> True
+            Aggregate_Column.Count_Nothing _ _ -> True
+            Aggregate_Column.Count_Not_Empty _ _ -> True
+            Aggregate_Column.Count_Empty _ _ -> True
+            Aggregate_Column.Percentile _ _ _ -> unsupported "Percentile"
+            Aggregate_Column.Mode _ _ -> unsupported "Mode"
+            Aggregate_Column.First _ _ _ _ -> unsupported "First"
+            Aggregate_Column.Last _ _ _ _ -> unsupported "Last"
+            Aggregate_Column.Maximum _ _ -> True
+            Aggregate_Column.Minimum _ _ -> True
+            Aggregate_Column.Shortest _ _ -> unsupported "Shortest"
+            Aggregate_Column.Longest _ _ -> unsupported "Longest"
+            Aggregate_Column.Standard_Deviation _ _ _ -> True
+            Aggregate_Column.Concatenate _ _ _ _ _ _ -> True
+            Aggregate_Column.Sum _ _ -> True
+            Aggregate_Column.Average _ _ -> True
+            Aggregate_Column.Median _ _ -> unsupported "Median"
 
     ## ---
        private: true

distribution/lib/Standard/Snowflake/0.0.0-dev/src/Internal/Snowflake_Dialect.enso

@@ -8,7 +8,6 @@ import Standard.Base.Runtime.Ref.Ref
 
 import Standard.Table.Internal.Problem_Builder.Problem_Builder
 from Standard.Table import Aggregate_Column, Column, Table, Value_Type
-from Standard.Table.Aggregate_Column.Aggregate_Column import all
 from Standard.Table.Errors import Inexact_Type_Coercion
 
 import Standard.Database.DB_Column.DB_Column

distribution/lib/Standard/Table/0.0.0-dev/docs/api/Aggregate_Column.md

@@ -18,6 +18,7 @@
     - Minimum column:(Standard.Base.Data.Text.Text|Standard.Base.Data.Numbers.Integer|Standard.Base.Any.Any)= as:Standard.Base.Data.Text.Text=
     - Mode column:(Standard.Base.Data.Text.Text|Standard.Base.Data.Numbers.Integer|Standard.Base.Any.Any)= as:Standard.Base.Data.Text.Text=
     - Percentile percentile:Standard.Base.Data.Numbers.Number= column:(Standard.Base.Data.Text.Text|Standard.Base.Data.Numbers.Integer|Standard.Base.Any.Any)= as:Standard.Base.Data.Text.Text=
+    - SQL function:Standard.Base.Data.Text.Text= arguments:(Standard.Base.Data.Vector.Vector Standard.Base.Any.Any)= as:Standard.Base.Data.Text.Text=
     - Shortest column:(Standard.Base.Data.Text.Text|Standard.Base.Data.Numbers.Integer|Standard.Base.Any.Any)= as:Standard.Base.Data.Text.Text=
     - Standard_Deviation column:(Standard.Base.Data.Text.Text|Standard.Base.Data.Numbers.Integer|Standard.Base.Any.Any)= as:Standard.Base.Data.Text.Text= population:Standard.Base.Data.Boolean.Boolean=
     - Sum column:(Standard.Base.Data.Text.Text|Standard.Base.Data.Numbers.Integer|Standard.Base.Any.Any)= as:Standard.Base.Data.Text.Text=

distribution/lib/Standard/Table/0.0.0-dev/src/Aggregate_Column.enso

@@ -198,3 +198,11 @@ type Aggregate_Column
           group longest value.
         - `as`: name of new column.
     Longest (column:Text|Integer|Any=0) (as:Text="") # Any needed because of 6866
+
+    ## Custom SQL aggregate function for database backends.
+
+       ## Arguments
+        - `function`: the SQL function to use for aggregation.
+        - `arguments`: arguments to the SQL function.
+        - `as`: name of new column.
+    SQL (function:Text=(Missing_Argument.throw "function")) (arguments:(Vector (Text | Sort_Column))=[]) (as:Text="")