Add DB_Table.offset for snowflake, postgres and sqlite (#12251)

* Red

* Fix tests

* Remove comment

* checkpoint

* Checkpoint

* 6 red

* 3 red

* Green

* Checkpoint

* Remove wrap around tests

* Green

* Refactor

* Adjust limit

* Clean up

* Refactor

* Refactor

* refactor

* Refactor

* Refactor

* Documentation

* Cleanup

* Comment

* Checkpoint

* Refactor

* fix

* More tests

* Table.new

* Cleanup

* changelog

* Fix test

* Code review changes

* Code review

* Postgres

* SQLite

* Snowflake

* Add missing aliases

* changelog

* Fix SQLite test

CHANGELOG.md

@@ -51,13 +51,15 @@
     `on_invalid_rows`. The default behaviour was also changed to add any extra
     columns instead of discarding them.
 - [Added DB_Table.Offset for SQLServer][12206]
+- [Added DB_Table.Offset for Snowflake, Postgres, SQLite][12251]
 
 [11926]: https://github.com/enso-org/enso/pull/11926
 [12031]: https://github.com/enso-org/enso/pull/12031
 [12071]: https://github.com/enso-org/enso/pull/12071
 [12092]: https://github.com/enso-org/enso/pull/12092
 [12231]: https://github.com/enso-org/enso/pull/12231
 [12206]: https://github.com/enso-org/enso/pull/12206
+[12251]: https://github.com/enso-org/enso/pull/12251
 
 #### Enso Language & Runtime
 

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

@@ -469,7 +469,8 @@ base_dialect_operations =
     contains = [["IS_IN", make_is_in], ["IS_IN_COLUMN", make_is_in_column]]
     types = [simple_cast]
     windows = [["ROW_NUMBER", make_row_number], ["ROW_NUMBER_IN_GROUP", make_row_number_in_group]]
-    base_dict = Dictionary.from_vector (arith + logic + compare + functions + agg + counts + text + nulls + contains + types + windows)
+    leadlag = [["LEAD", _make_lead_lag "LEAD"], ["LAG", _make_lead_lag "LAG"], ["LEAD_CLOSEST", _make_lead_lag_closest_value "LEAD"], ["LAG_CLOSEST", _make_lead_lag_closest_value "LAG"]]
+    base_dict = Dictionary.from_vector (arith + logic + compare + functions + agg + counts + text + nulls + contains + types + windows + leadlag)
     Dialect_Operations.Value base_dict
 
 ## PRIVATE
@@ -740,6 +741,38 @@ default_fetch_types_query dialect expression context where_filter_always_false_l
 ## PRIVATE
 default_generate_collate collation_name:Text quote_char:Text='"' -> Text = ' COLLATE ' + quote_char + collation_name + quote_char
 
+_build_partition_sql grouping:SQL_Builder ordering:SQL_Builder -> SQL_Builder =
+    group_part = if grouping.is_empty then "" else
+        SQL_Builder.code "PARTITION BY " ++ grouping
+    SQL_Builder.code "OVER(" ++ group_part ++ " ORDER BY " ++ ordering ++ ")"
+
+## PRIVATE
+_build_lead_lag_sql lead_lag:Text n:SQL_Builder colName:SQL_Builder grouping:SQL_Builder ordering:SQL_Builder -> SQL_Builder =
+    partition_sql = _build_partition_sql grouping ordering
+    SQL_Builder.code "(" ++ lead_lag ++ "(" ++ colName ++ ", " ++ n ++ ", NULL) " ++ partition_sql ++ ")"
+
+## PRIVATE
+_make_lead_lag lead_lag:Text arguments:Vector -> SQL_Builder = if arguments.length != 4 then Error.throw (Illegal_State.Error "Wrong amount of parameters in LEAD/LAG IR. This is a bug in the Database library.") else
+    n = arguments.at 0
+    colName = arguments.at 1
+    grouping = arguments.at 2
+    ordering = arguments.at 3
+    _build_lead_lag_sql lead_lag n colName grouping ordering
+
+## PRIVATE
+_make_lead_lag_closest_value lead_lag:Text arguments:Vector -> SQL_Builder = if arguments.length != 5 then Error.throw (Illegal_State.Error "Wrong amount of parameters in LEAD/LAG IR. This is a bug in the Database library.") else
+    n = arguments.at 0
+    colName = arguments.at 1
+    grouping = arguments.at 2
+    ordering_for_lead_lag = arguments.at 3
+    ordering_for_row_number = arguments.at 4
+
+    lead_lag_sql = _build_lead_lag_sql lead_lag n colName grouping ordering_for_lead_lag     
+    partition_sql_for_row_number = _build_partition_sql grouping ordering_for_row_number
+    fill_sql = SQL_Builder.code "FIRST_VALUE(" ++ colName ++ ") " ++ partition_sql_for_row_number
+    SQL_Builder.code "CASE WHEN ROW_NUMBER() " ++ partition_sql_for_row_number ++ " <= " ++ n ++ " THEN " ++ fill_sql ++ " ELSE " ++ lead_lag_sql ++ " END"
+
+
 ## PRIVATE
    Helper class for shortening the binder names generated for WITH clauses.
 

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

@@ -269,7 +269,6 @@ type Postgres_Dialect
        Checks if a feature is supported by the dialect.
     is_feature_supported self feature:Feature -> Boolean =
         case feature of
-            Feature.Offset -> False
             _ -> True
 
     ## PRIVATE

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

@@ -278,7 +278,6 @@ type SQLite_Dialect
        Checks if a feature is supported by the dialect.
     is_feature_supported self feature:Feature -> Boolean =
         case feature of
-            Feature.Offset -> False
             _ -> True
 
     ## PRIVATE

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

@@ -214,7 +214,7 @@ operations_dict =
     always_integer_ops = ["COUNT", "COUNT_IS_NULL", "COUNT_DISTINCT", "COUNT_DISTINCT_INCLUDE_NULL", "COUNT_EMPTY", "COUNT_NOT_EMPTY", "COUNT_ROWS", "COUNT_OVER_PARTITION", "ROW_NUMBER", "ROW_NUMBER_IN_GROUP", "LENGTH"]
     same_as_first = ["TRUNCATE", "CEIL", "FLOOR", "FIRST", "LAST"]
     arithmetic_ops = ["ADD_NUMBER", "-", "*", "^", "%", "SUM"]
-    merge_input_types_ops = ["ROW_MAX", "ROW_MIN", "MAX", "MIN", "FILL_NULL", "COALESCE"]
+    merge_input_types_ops = ["ROW_MAX", "ROW_MIN", "MAX", "MIN", "FILL_NULL", "COALESCE", "LEAD", "LAG", "LEAD_CLOSEST", "LAG_CLOSEST"]
     others = [["IIF", handle_iif], ["CAST", handle_cast], ["CASE", handle_case], ["RUNTIME_ERROR", handle_runtime_error]]
     Dictionary.from_vector <|
         v1 = always_boolean_ops.map [_, const SQLite_Types.boolean]

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

@@ -257,7 +257,6 @@ type Snowflake_Dialect
        Checks if a feature is supported by the dialect.
     is_feature_supported self feature:Feature -> Boolean =
         case feature of
-            Feature.Offset -> False
             _ -> True
 
     ## PRIVATE

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

@@ -3160,7 +3160,7 @@ type Table
     union self tables:(Table | Vector) (columns_to_keep : Columns_To_Keep = ..In_Any_Warn_On_Missing) (match_columns : Match_Columns = ..By_Name) (on_problems : Problem_Behavior = ..Report_Warning) =
         Table.from_union ([self] + Vector.unify_vector_or_element tables) columns_to_keep match_columns on_problems
 
-    ## ALIAS drop_missing_rows, dropna
+    ## ALIAS drop_empty_rows, drop_missing_rows, dropna, filter_empty_rows, remove_blank_rows, remove_empty_rows, remove_missing_rows
        GROUP Standard.Base.Selections
        ICON preparation
        Remove rows which are all blank or containing blank values.

test/Table_Tests/src/Common_Table_Operations/Util.enso

@@ -82,8 +82,7 @@ Error.should_equal_tz_agnostic self other =
 ## PRIVATE
    Builds a table ensuring that the rows are in the order as given.
 build_sorted_table setup table_structure =
-    # Workaround for https://github.com/enso-org/enso/issues/10321
-    if setup.prefix.contains "Snowflake" . not && setup.prefix.contains "SQLServer" . not then setup.table_builder table_structure else
+    if setup.is_database . not then setup.table_builder table_structure else
         row_count = case table_structure.first of
             def : Vector -> def.second.length
             col : Column -> col.length