feat: Simplify NOT(IN ..) to NOT IN and NOT (EXISTS ..) to NOT EXISTS

datafusion/expr/src/operation.rs

@@ -17,8 +17,8 @@
 
 //! This module contains implementations of operations (unary, binary etc.) for DataFusion expressions.
 
+use crate::expr::{Exists, Expr, InList, InSubquery, Like};
 use crate::expr_fn::binary_expr;
-use crate::{Expr, Like};
 use datafusion_expr_common::operator::Operator;
 use std::ops::{self, Not};
 
@@ -153,6 +153,19 @@ impl Not for Expr {
                 escape_char,
                 case_insensitive,
             )),
+            Expr::InList(InList {
+                expr,
+                list,
+                negated,
+            }) => Expr::InList(InList::new(expr, list, !negated)),
+            Expr::Exists(Exists { subquery, negated }) => {
+                Expr::Exists(Exists::new(subquery, !negated))
+            }
+            Expr::InSubquery(InSubquery {
+                expr,
+                subquery,
+                negated,
+            }) => Expr::InSubquery(InSubquery::new(expr, subquery, !negated)),
             _ => Expr::Not(Box::new(self)),
         }
     }

datafusion/optimizer/src/simplify_expressions/simplify_exprs.rs

@@ -926,4 +926,152 @@ mod tests {
         "#
         )
     }
+
+    #[test]
+    fn simplify_not_in_list() -> Result<()> {
+        let schema = Schema::new(vec![Field::new("a", DataType::Utf8, false)]);
+        let table_scan = table_scan(Some("test"), &schema, None)?.build()?;
+
+        let plan = LogicalPlanBuilder::from(table_scan)
+            .filter(col("a").in_list(vec![lit("a"), lit("b")], false).not())?
+            .build()?;
+
+        assert_optimized_plan_equal!(
+            plan,
+            @ r#"
+        Filter: test.a != Utf8("a") AND test.a != Utf8("b")
+          TableScan: test
+        "#
+        )
+    }
+
+    #[test]
+    fn simplify_not_not_in_list() -> Result<()> {
+        let schema = Schema::new(vec![Field::new("a", DataType::Utf8, false)]);
+        let table_scan = table_scan(Some("test"), &schema, None)?.build()?;
+
+        let plan = LogicalPlanBuilder::from(table_scan)
+            .filter(
+                col("a")
+                    .in_list(vec![lit("a"), lit("b")], false)
+                    .not()
+                    .not(),
+            )?
+            .build()?;
+
+        assert_optimized_plan_equal!(
+            plan,
+            @ r#"
+        Filter: test.a = Utf8("a") OR test.a = Utf8("b")
+          TableScan: test
+        "#
+        )
+    }
+
+    #[test]
+    fn simplify_not_exists() -> Result<()> {
+        let schema = Schema::new(vec![Field::new("a", DataType::Utf8, false)]);
+        let table_scan = table_scan(Some("test"), &schema, None)?.build()?;
+        let table_scan2 =
+            datafusion_expr::table_scan(Some("test2"), &schema, None)?.build()?;
+
+        let plan = LogicalPlanBuilder::from(table_scan)
+            .filter(
+                exists(Arc::new(LogicalPlanBuilder::from(table_scan2).build()?)).not(),
+            )?
+            .build()?;
+
+        assert_optimized_plan_equal!(
+            plan,
+            @ r"
+        Filter: NOT EXISTS (<subquery>)
+          Subquery:
+            TableScan: test2
+          TableScan: test
+        "
+        )
+    }
+
+    #[test]
+    fn simplify_not_not_exists() -> Result<()> {
+        let schema = Schema::new(vec![Field::new("a", DataType::Utf8, false)]);
+        let table_scan = table_scan(Some("test"), &schema, None)?.build()?;
+        let table_scan2 =
+            datafusion_expr::table_scan(Some("test2"), &schema, None)?.build()?;
+
+        let plan = LogicalPlanBuilder::from(table_scan)
+            .filter(
+                exists(Arc::new(LogicalPlanBuilder::from(table_scan2).build()?))
+                    .not()
+                    .not(),
+            )?
+            .build()?;
+
+        assert_optimized_plan_equal!(
+            plan,
+            @ r"
+        Filter: EXISTS (<subquery>)
+          Subquery:
+            TableScan: test2
+          TableScan: test
+        "
+        )
+    }
+
+    #[test]
+    fn simplify_not_in_subquery() -> Result<()> {
+        let schema = Schema::new(vec![Field::new("a", DataType::Utf8, false)]);
+        let table_scan = table_scan(Some("test"), &schema, None)?.build()?;
+        let table_scan2 =
+            datafusion_expr::table_scan(Some("test2"), &schema, None)?.build()?;
+
+        let plan = LogicalPlanBuilder::from(table_scan)
+            .filter(
+                in_subquery(
+                    col("a"),
+                    Arc::new(LogicalPlanBuilder::from(table_scan2).build()?),
+                )
+                .not(),
+            )?
+            .build()?;
+
+        assert_optimized_plan_equal!(
+            plan,
+            @ r"
+        Filter: test.a NOT IN (<subquery>)
+          Subquery:
+            TableScan: test2
+          TableScan: test
+        "
+        )
+    }
+
+    #[test]
+    fn simplify_not_not_in_subquery() -> Result<()> {
+        let schema = Schema::new(vec![Field::new("a", DataType::Utf8, false)]);
+        let table_scan = table_scan(Some("test"), &schema, None)?.build()?;
+        let table_scan2 =
+            datafusion_expr::table_scan(Some("test2"), &schema, None)?.build()?;
+
+        let plan = LogicalPlanBuilder::from(table_scan)
+            .filter(
+                in_subquery(
+                    col("a"),
+                    Arc::new(LogicalPlanBuilder::from(table_scan2).build()?),
+                )
+                .not()
+                .not(),
+            )?
+            .build()?;
+
+        assert_optimized_plan_equal!(
+            plan,
+            @ r"
+        Filter: test.a IN (<subquery>)
+          Subquery:
+            TableScan: test2
+          TableScan: test
+        "
+        )
+    }
 }