Implement cast for Table and Column (#6711)

Closes #6112

CHANGELOG.md

@@ -445,6 +445,7 @@
 - [Added `at_least_one` flag to `Table.tokenize_to_rows`.][6539]
 - [Moved `Redshift` connector into a separate `AWS` library.][6550]
 - [Added `Date_Range`.][6621]
+- [Implemented the `cast` operation for `Table` and `Column`.][6711]
 
 [debug-shortcuts]:
   https://github.com/enso-org/enso/blob/develop/app/gui/docs/product/shortcuts.md#debug
@@ -655,6 +656,7 @@
 [6539]: https://github.com/enso-org/enso/pull/6539
 [6550]: https://github.com/enso-org/enso/pull/6550
 [6621]: https://github.com/enso-org/enso/pull/6621
+[6711]: https://github.com/enso-org/enso/pull/6711
 
 #### Enso Compiler
 

distribution/lib/Standard/Database/0.0.0-dev/src/Data/Column.enso

@@ -9,7 +9,8 @@ import Standard.Table.Internal.Java_Problems
 import Standard.Table.Internal.Problem_Builder.Problem_Builder
 import Standard.Table.Internal.Widget_Helpers
 from Standard.Table import Sort_Column, Data_Formatter, Value_Type, Auto
-from Standard.Table.Errors import Floating_Point_Equality, Inexact_Type_Coercion, Invalid_Value_Type, Lossy_Conversion
+from Standard.Table.Errors import Floating_Point_Equality, Inexact_Type_Coercion, Invalid_Value_Type, Conversion_Failure
+from Standard.Table.Internal.Cast_Helpers import check_cast_compatibility
 
 import project.Connection.Connection.Connection
 import project.Data.SQL_Statement.SQL_Statement
@@ -1002,18 +1003,13 @@ type Column
         _ = [format, locale]
         Error.throw <| Unsupported_Database_Operation.Error "`Column.format` is not implemented yet for the Database backends."
 
-    ## PRIVATE
-       UNSTABLE
-       Cast the column to a specific type.
+    ## Cast the column to a specific type.
 
        Arguments:
        - value_type: The `Value_Type` to cast the column to.
        - on_problems: Specifies how to handle problems if they occur, reporting
          them as warnings by default.
 
-       TODO [RW] this is a prototype needed for debugging, proper implementation
-       and testing will come with #6112.
-
        In the Database backend, this will boil down to a CAST operation.
        In the in-memory backend, a conversion will be performed according to
        the following rules:
@@ -1024,6 +1020,9 @@ type Column
          length.
        - Conversion between numeric types will replace values exceeding the
          range of the target type with `Nothing`.
+       - Converting decimal numbers into integers will truncate or round them,
+         depending on the backend. If more control is needed, use the various
+         rounding functions (such as `round` or `floor`).
        - Booleans may also be converted to numbers, with `True` being converted
          to `1` and `False` to `0`. The reverse is not supported - use `iif`
          instead.
@@ -1032,32 +1031,27 @@ type Column
        - If a `Date` is to be converted to `Date_Time`, it will be set at
          midnight of the default system timezone.
 
-       ? Conversion Precision
-
-         In the in-memory backend, if the conversion is lossy, a
-         `Lossy_Conversion` warning will be reported. The only exception is when
-         truncating a column which is already a text column - as then the
-         truncation seems like an intended behaviour, so it is not reported. If
-         truncating needs to occur when converting a non-text column, a warning
-         will still be reported.
-
-         Currently, the warning is not reported for Database backends.
+        If the target type cannot fit some of the values (for example due to too
+        small range), a `Conversion_Failure` may be reported according to the
+        `on_problems` rules. The Database backends may fail with `SQL_Error`
+        instead.
 
        ? Inexact Target Type
 
          If the backend does not support the requested target type, the closest
          supported type is chosen and a `Inexact_Type_Coercion` problem is
          reported.
-    cast : Value_Type -> Problem_Behavior -> Column ! Illegal_Argument | Inexact_Type_Coercion | Lossy_Conversion
-    cast self value_type=self.value_type on_problems=Problem_Behavior.Report_Warning =
-        dialect = self.connection.dialect
-        type_mapping = dialect.get_type_mapping
-        target_sql_type = type_mapping.value_type_to_sql value_type on_problems
-        target_sql_type.if_not_error <|
-            infer_from_database new_expression =
-                SQL_Type_Reference.new self.connection self.context new_expression
-            new_column = dialect.make_cast self.as_internal target_sql_type infer_from_database
-            Column.Value new_column.name self.connection new_column.sql_type_reference new_column.expression self.context
+    cast : Value_Type -> Problem_Behavior -> Column ! Illegal_Argument | Inexact_Type_Coercion | Conversion_Failure
+    cast self value_type on_problems=Problem_Behavior.Report_Warning =
+        check_cast_compatibility self.value_type value_type <|
+            dialect = self.connection.dialect
+            type_mapping = dialect.get_type_mapping
+            target_sql_type = type_mapping.value_type_to_sql value_type on_problems
+            target_sql_type.if_not_error <|
+                infer_from_database new_expression =
+                    SQL_Type_Reference.new self.connection self.context new_expression
+                new_column = dialect.make_cast self.as_internal target_sql_type infer_from_database
+                Column.Value new_column.name self.connection new_column.sql_type_reference new_column.expression self.context
 
     ## ALIAS Transform Column
 

distribution/lib/Standard/Database/0.0.0-dev/src/Data/Table.enso

@@ -1526,9 +1526,7 @@ type Table
         _ = [column, pattern, case_sensitivity, parse_values, on_problems]
         Error.throw (Unsupported_Database_Operation.Error "Table.parse_to_columns is not implemented yet for the Database backends.")
 
-    ## PRIVATE
-       UNSTABLE
-       Cast the selected columns to a specific type.
+    ## Cast the selected columns to a specific type.
 
        Returns a new table in which the selected columns are replaced with
        columns having the new types.
@@ -1539,9 +1537,6 @@ type Table
        - on_problems: Specifies how to handle problems if they occur, reporting
          them as warnings by default.
 
-       TODO [RW] this is a prototype needed for debugging, proper implementation
-       and testing will come with #6112.
-
        In the Database backend, this will boil down to a CAST operation.
        In the in-memory backend, a conversion will be performed according to
        the following rules:
@@ -1552,6 +1547,9 @@ type Table
          length.
        - Conversion between numeric types will replace values exceeding the
          range of the target type with `Nothing`.
+       - Converting decimal numbers into integers will truncate or round them,
+         depending on the backend. If more control is needed, use the various
+         rounding functions (such as `round` or `floor`).
        - Booleans may also be converted to numbers, with `True` being converted
          to `1` and `False` to `0`. The reverse is not supported - use `iif`
          instead.
@@ -1560,27 +1558,21 @@ type Table
        - If a `Date` is to be converted to `Date_Time`, it will be set at
          midnight of the default system timezone.
 
-       ? Conversion Precision
-
-         In the in-memory backend, if the conversion is lossy, a
-         `Lossy_Conversion` warning will be reported. The only exception is when
-         truncating a column which is already a text column - as then the
-         truncation seems like an intended behaviour, so it is not reported. If
-         truncating needs to occur when converting a non-text column, a warning
-         will still be reported.
-
-         Currently, the warning is not reported for Database backends.
+        If the target type cannot fit some of the values (for example due to too
+        small range), a `Conversion_Failure` may be reported according to the
+        `on_problems` rules. The Database backends may fail with `SQL_Error`
+        instead.
 
        ? Inexact Target Type
 
          If the backend does not support the requested target type, the closest
          supported type is chosen and a `Inexact_Type_Coercion` problem is
          reported.
     @columns Widget_Helpers.make_column_name_vector_selector
-    cast : (Text | Integer | Column_Selector | Vector (Integer | Text | Column_Selector)) -> Value_Type -> Problem_Behavior -> Table ! Illegal_Argument | Inexact_Type_Coercion | Lossy_Conversion
-    cast self columns=[0] value_type=Value_Type.Char on_problems=Problem_Behavior.Report_Warning =
-        selected = self.select_columns columns
-        selected.columns.fold self table-> column_to_cast->
+    cast : (Text | Integer | Column_Selector | Vector (Integer | Text | Column_Selector)) -> Value_Type -> Boolean -> Problem_Behavior -> Table ! Illegal_Argument | Inexact_Type_Coercion | Conversion_Failure
+    cast self columns=[0] value_type error_on_missing_columns=True on_problems=Problem_Behavior.Report_Warning =
+        selected = self.columns_helper.resolve_columns columns error_on_missing_columns=error_on_missing_columns on_problems=on_problems
+        selected.fold self table-> column_to_cast->
             new_column = column_to_cast.cast value_type on_problems
             table.set new_column new_name=column_to_cast.name set_mode=Set_Mode.Update
 

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

@@ -145,12 +145,19 @@ type SQLite_Dialect
     make_cast : Internal_Column -> SQL_Type -> (SQL_Expression -> SQL_Type_Reference) -> Internal_Column
     make_cast self column target_type _ =
         mapping = self.get_type_mapping
-        sql_type_text = mapping.sql_type_to_text target_type
-        new_expression = SQL_Expression.Operation "CAST" [column.expression, SQL_Expression.Literal sql_type_text]
-        # We override the type here, because SQLite gets it wrong if the column starts with NULL values.
+        target_value_type = mapping.sql_type_to_value_type target_type
+        custom_cast = make_custom_cast column target_value_type mapping
+        new_expression = custom_cast.if_nothing <|
+            self.make_cast_expression column target_type
         new_sql_type_reference = SQL_Type_Reference.from_constant target_type
         Internal_Column.Value column.name new_sql_type_reference new_expression
 
+    ## PRIVATE
+    make_cast_expression self column target_type =
+        mapping = self.get_type_mapping
+        sql_type_text = mapping.sql_type_to_text target_type
+        SQL_Expression.Operation "CAST" [column.expression, SQL_Expression.Literal sql_type_text]
+
     ## PRIVATE
     needs_execute_query_for_type_inference : Boolean
     needs_execute_query_for_type_inference self = True
@@ -164,12 +171,15 @@ type SQLite_Dialect
        So after unifying columns with mixed types, we add a cast to ensure that.
     adapt_unified_column : Internal_Column -> Value_Type -> (SQL_Expression -> SQL_Type_Reference) -> Internal_Column
     adapt_unified_column self column approximate_result_type infer_result_type_from_database_callback =
+        _ = infer_result_type_from_database_callback
         # TODO [RW] This may be revisited with #6281.
         case approximate_result_type of
             Nothing -> column
             _ ->
                 sql_type = self.get_type_mapping.value_type_to_sql approximate_result_type Problem_Behavior.Ignore
-                self.make_cast column sql_type infer_result_type_from_database_callback
+                new_expression = self.make_cast_expression column sql_type
+                new_sql_type_reference = SQL_Type_Reference.from_constant sql_type
+                Internal_Column.Value column.name new_sql_type_reference new_expression
 
     ## PRIVATE
     prepare_fetch_types_query : SQL_Expression -> Context -> SQL_Statement
@@ -353,3 +363,11 @@ decimal_div = Base_Generator.lift_binary_op "/" x-> y->
 ## PRIVATE
 mod_op = Base_Generator.lift_binary_op "mod" x-> y->
     x ++ " - FLOOR(CAST(" ++ x ++ " AS REAL) / CAST(" ++ y ++ " AS REAL)) * " ++ y
+
+## PRIVATE
+   It will return `Nothing` if the type does not require custom logic.
+make_custom_cast column target_value_type type_mapping =
+    if target_value_type.is_text then
+        column_type = type_mapping.sql_type_to_value_type column.sql_type_reference.get
+        if column_type == Value_Type.Boolean then
+            SQL_Expression.Operation "IIF" [column.expression, SQL_Expression.Literal "'true'", SQL_Expression.Literal "'false'"]

distribution/lib/Standard/Table/0.0.0-dev/src/Data/Column.enso

@@ -13,6 +13,7 @@ import project.Data.Type.Enso_Types
 import project.Data.Type.Storage
 import project.Data.Type.Value_Type_Helpers
 import project.Data.Table.Table
+import project.Internal.Cast_Helpers
 import project.Internal.Java_Problems
 import project.Internal.Naming_Helpers.Naming_Helpers
 import project.Internal.Parse_Values_Helper
@@ -21,7 +22,7 @@ import project.Data.Type.Value_Type_Helpers
 
 from project.Data.Table import print_table
 from project.Data.Type.Value_Type import Value_Type, Auto
-from project.Errors import No_Index_Set_Error, Floating_Point_Equality, Invalid_Value_Type, Inexact_Type_Coercion
+from project.Errors import No_Index_Set_Error, Floating_Point_Equality, Invalid_Value_Type, Inexact_Type_Coercion, Conversion_Failure
 from project.Internal.Java_Exports import make_string_builder
 
 polyglot java import org.enso.table.data.column.operation.map.MapOperationProblemBuilder
@@ -1279,6 +1280,54 @@ type Column
             _ -> Error.throw <| Illegal_Argument.Error <| "Unsupported format type: " + format.to_text
         new_column
 
+    ## Cast the column to a specific type.
+
+       Arguments:
+       - value_type: The `Value_Type` to cast the column to.
+       - on_problems: Specifies how to handle problems if they occur, reporting
+         them as warnings by default.
+
+       In the Database backend, this will boil down to a CAST operation.
+       In the in-memory backend, a conversion will be performed according to
+       the following rules:
+       - Anything can be cast into the `Mixed` type.
+       - Converting to a `Char` type, the elements of the column will be
+         converted to text. If it is fixed length, the texts will be trimmed or
+         padded on the right with the space character to match the desired
+         length.
+       - Conversion between numeric types will replace values exceeding the
+         range of the target type with `Nothing`.
+       - Converting decimal numbers into integers will truncate or round them,
+         depending on the backend. If more control is needed, use the various
+         rounding functions (such as `round` or `floor`).
+       - Booleans may also be converted to numbers, with `True` being converted
+         to `1` and `False` to `0`. The reverse is not supported - use `iif`
+         instead.
+       - A `Date_Time` may be converted into a `Date` or `Time` type - the
+         resulting value will be truncated to the desired type.
+       - If a `Date` is to be converted to `Date_Time`, it will be set at
+         midnight of the default system timezone.
+
+        If the target type cannot fit some of the values (for example due to too
+        small range), a `Conversion_Failure` may be reported according to the
+        `on_problems` rules. The Database backends may fail with `SQL_Error`
+        instead.
+
+       ? Inexact Target Type
+
+         If the backend does not support the requested target type, the closest
+         supported type is chosen and a `Inexact_Type_Coercion` problem is
+         reported.
+    cast : Value_Type -> Problem_Behavior -> Column ! Illegal_Argument | Inexact_Type_Coercion | Conversion_Failure
+    cast self value_type on_problems=Problem_Behavior.Report_Warning =
+        Cast_Helpers.check_cast_compatibility self.value_type value_type <|
+            target_storage_type = Storage.from_value_type value_type on_problems
+            cast_problem_builder = Cast_Helpers.new_java_problem_builder self.name value_type
+            new_storage = self.java_column.getStorage.cast target_storage_type cast_problem_builder.to_java
+            problems = cast_problem_builder.get_problems
+            on_problems.attach_problems_before problems <|
+                Column.from_storage self.name new_storage
+
     ## ALIAS Transform Column
 
        Applies `function` to each item in this column and returns the column

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

@@ -873,9 +873,7 @@ type Table
             parse_problem_builder.attach_problems_before on_problems <|
                 Table.new new_columns
 
-    ## PRIVATE
-       UNSTABLE
-       Cast the selected columns to a specific type.
+    ## Cast the selected columns to a specific type.
 
        Returns a new table in which the selected columns are replaced with
        columns having the new types.
@@ -886,9 +884,6 @@ type Table
        - on_problems: Specifies how to handle problems if they occur, reporting
          them as warnings by default.
 
-       TODO [RW] this is a prototype needed for debugging, proper implementation
-       and testing will come with #6112.
-
        In the Database backend, this will boil down to a CAST operation.
        In the in-memory backend, a conversion will be performed according to
        the following rules:
@@ -899,6 +894,9 @@ type Table
          length.
        - Conversion between numeric types will replace values exceeding the
          range of the target type with `Nothing`.
+       - Converting decimal numbers into integers will truncate or round them,
+         depending on the backend. If more control is needed, use the various
+         rounding functions (such as `round` or `floor`).
        - Booleans may also be converted to numbers, with `True` being converted
          to `1` and `False` to `0`. The reverse is not supported - use `iif`
          instead.
@@ -907,28 +905,23 @@ type Table
        - If a `Date` is to be converted to `Date_Time`, it will be set at
          midnight of the default system timezone.
 
-       ? Conversion Precision
-
-         In the in-memory backend, if the conversion is lossy, a
-         `Lossy_Conversion` warning will be reported. The only exception is when
-         truncating a column which is already a text column - as then the
-         truncation seems like an intended behaviour, so it is not reported. If
-         truncating needs to occur when converting a non-text column, a warning
-         will still be reported.
-
-         Currently, the warning is not reported for Database backends.
+        If the target type cannot fit some of the values (for example due to too
+        small range), a `Conversion_Failure` may be reported according to the
+        `on_problems` rules. The Database backends may fail with `SQL_Error`
+        instead.
 
        ? Inexact Target Type
 
          If the backend does not support the requested target type, the closest
          supported type is chosen and a `Inexact_Type_Coercion` problem is
          reported.
     @columns Widget_Helpers.make_column_name_vector_selector
-    cast : (Text | Integer | Column_Selector | Vector (Integer | Text | Column_Selector)) -> Value_Type -> Problem_Behavior -> Table ! Illegal_Argument | Inexact_Type_Coercion | Lossy_Conversion
-    cast self columns=[0] value_type=Value_Type.Char on_problems=Problem_Behavior.Report_Warning =
-        _ = [columns, value_type, on_problems]
-        ## TODO [RW] actual implementation in #6112
-        self
+    cast : (Text | Integer | Column_Selector | Vector (Integer | Text | Column_Selector)) -> Value_Type -> Boolean -> Problem_Behavior -> Table ! Illegal_Argument | Inexact_Type_Coercion | Conversion_Failure
+    cast self columns=[0] value_type error_on_missing_columns=True on_problems=Problem_Behavior.Report_Warning =
+        selected = self.columns_helper.resolve_columns columns error_on_missing_columns=error_on_missing_columns on_problems=on_problems
+        selected.fold self table-> column_to_cast->
+            new_column = column_to_cast.cast value_type on_problems
+            table.set new_column new_name=column_to_cast.name set_mode=Set_Mode.Update
 
     ## Splits a column of text into a set of new columns.
        The original column will be removed from the table.