fix: Optimize !~ '.*' case to col IS NULL AND Boolean(NULL) instead of Eq "" (#20702)

## Which issue does this PR close?

<!--
We generally require a GitHub issue to be filed for all bug fixes and
enhancements and this helps us generate change logs for our releases.
You can link an issue to this PR using the GitHub syntax. For example
`Closes #123` indicates that this PR will close issue #123.
-->

- Closes #20701

## Rationale for this change

<!--
Why are you proposing this change? If this is already explained clearly
in the issue then this section is not needed.
Explaining clearly why changes are proposed helps reviewers understand
your changes and offer better suggestions for fixes.
-->


## What changes are included in this PR?

<!--
There is no need to duplicate the description in the issue here but it
is sometimes worth providing a summary of the individual changes in this
PR.
-->
A pre-existing optimization rule for the `!~ .*` (regexp not match) case
rewrote the plan to `Eq ""`, which would return empty strings as part of
the result. This is incorrect and doesn't match the output without the
optimization rule.

Instead, this PR rewrites the plan to simply `col IS NULL AND
Boolean(NULL)` or, in other words, "NULL if col is NULL else false."

I've confirmed this behavior matches the result of running queries
manually with the optimization rule turned off.

## Are these changes tested?

<!--
We typically require tests for all PRs in order to:
1. Prevent the code from being accidentally broken by subsequent changes
2. Serve as another way to document the expected behavior of the code

If tests are not included in your PR, please explain why (for example,
are they covered by existing tests)?
-->
Fixed expected output in tests. Added new tests for nulls

## Are there any user-facing changes?

<!--
If there are user-facing changes then we may require documentation to be
updated before approving the PR.
-->

<!--
If there are any breaking changes to public APIs, please add the `api
change` label.
-->

Yes, a minor bug fix. When querying `s !~ .*`, empty strings will no
longer be included in the result which is consistent with the behavior
without the optimization rule.

---------

Co-authored-by: Andrew Lamb <andrew@nerdnetworks.org>

Author: petern48

datafusion/optimizer/src/simplify_expressions/regex.rs

@@ -16,7 +16,7 @@
 // under the License.
 
 use datafusion_common::tree_node::Transformed;
-use datafusion_common::{DataFusionError, Result};
+use datafusion_common::{DataFusionError, Result, ScalarValue};
 use datafusion_expr::{BinaryExpr, Expr, Like, Operator, lit};
 use regex_syntax::hir::{Capture, Hir, HirKind, Literal, Look};
 
@@ -39,7 +39,7 @@ const ANY_CHAR_REGEX_PATTERN: &str = ".*";
 /// - partial anchored regex patterns (e.g. `^foo`) to `LIKE 'foo%'`
 /// - combinations (alternatives) of the above, will be concatenated with `OR` or `AND`
 /// - `EQ .*` to NotNull
-/// - `NE .*` means IS EMPTY
+/// - `NE .*` to col IS NULL AND Boolean(NULL) (false for any string, or NULL if col is NULL)
 ///
 /// Dev note: unit tests of this function are in `expr_simplifier.rs`, case `test_simplify_regex`.
 pub fn simplify_regex_expr(
@@ -68,12 +68,11 @@ pub fn simplify_regex_expr(
     // Handle the special case for ".*" pattern
     if pattern == ANY_CHAR_REGEX_PATTERN {
         let new_expr = if mode.not {
-            // not empty
-            let empty_lit = Box::new(string_scalar.to_expr(""));
+            let null_bool = lit(ScalarValue::Boolean(None));
             Expr::BinaryExpr(BinaryExpr {
-                left,
-                op: Operator::Eq,
-                right: empty_lit,
+                left: Box::new(left.is_null()),
+                op: Operator::And,
+                right: Box::new(null_bool),
             })
         } else {
             // not null

datafusion/optimizer/src/simplify_expressions/simplify_exprs.rs

@@ -155,6 +155,7 @@ mod tests {
     use arrow::datatypes::{DataType, Field, Schema};
     use chrono::{DateTime, Utc};
 
+    use datafusion_common::ScalarValue;
     use datafusion_expr::logical_plan::builder::table_scan_with_filters;
     use datafusion_expr::logical_plan::table_scan;
     use datafusion_expr::*;
@@ -870,7 +871,7 @@ mod tests {
         ]);
         let table_scan = table_scan(Some("test"), &schema, None)?.build()?;
 
-        // Test `= ".*"` transforms to true (except for empty strings)
+        // Test `~ ".*"` transforms to true for any non-NULL string
         let plan = LogicalPlanBuilder::from(table_scan.clone())
             .filter(binary_expr(col("a"), Operator::RegexMatch, lit(".*")))?
             .build()?;
@@ -883,22 +884,22 @@ mod tests {
         "
         )?;
 
-        // Test `!= ".*"` transforms to checking if the column is empty
+        // Test `!~ ".*"` preserves NULL semantics while remaining false for non-NULL strings
         let plan = LogicalPlanBuilder::from(table_scan.clone())
             .filter(binary_expr(col("a"), Operator::RegexNotMatch, lit(".*")))?
             .build()?;
 
         assert_optimized_plan_equal!(
             plan,
-            @ r#"
-        Filter: test.a = Utf8("")
+            @ r"
+        Filter: test.a IS NULL AND Boolean(NULL)
           TableScan: test
-        "#
+        "
         )?;
 
         // Test case-insensitive versions
 
-        // Test `=~ ".*"` (case-insensitive) transforms to true (except for empty strings)
+        // Test `~* ".*"` transforms to true for any non-NULL string
         let plan = LogicalPlanBuilder::from(table_scan.clone())
             .filter(binary_expr(col("b"), Operator::RegexIMatch, lit(".*")))?
             .build()?;
@@ -911,17 +912,51 @@ mod tests {
         "
         )?;
 
-        // Test `!~ ".*"` (case-insensitive) transforms to checking if the column is empty
+        // Test NULL `!~ ".*"` transforms to Boolean(NULL)
+        let plan = LogicalPlanBuilder::from(table_scan.clone())
+            .filter(binary_expr(
+                lit(ScalarValue::Utf8(None)),
+                Operator::RegexNotMatch,
+                lit(".*"),
+            ))?
+            .build()?;
+
+        assert_optimized_plan_equal!(
+            plan,
+            @ r"
+        Filter: Boolean(NULL)
+          TableScan: test
+        "
+        )?;
+
+        // Test `!~* ".*"` preserves NULL semantics while remaining false for non-NULL strings
         let plan = LogicalPlanBuilder::from(table_scan.clone())
             .filter(binary_expr(col("a"), Operator::RegexNotIMatch, lit(".*")))?
             .build()?;
 
         assert_optimized_plan_equal!(
             plan,
-            @ r#"
-        Filter: test.a = Utf8("")
+            @ r"
+        Filter: test.a IS NULL AND Boolean(NULL)
           TableScan: test
-        "#
+        "
+        )?;
+
+        // Test NULL `!~* ".*"` transforms to Boolean(NULL)
+        let plan = LogicalPlanBuilder::from(table_scan.clone())
+            .filter(binary_expr(
+                lit(ScalarValue::Utf8(None)),
+                Operator::RegexNotIMatch,
+                lit(".*"),
+            ))?
+            .build()?;
+
+        assert_optimized_plan_equal!(
+            plan,
+            @ r"
+        Filter: Boolean(NULL)
+          TableScan: test
+        "
         )
     }
 

datafusion/sqllogictest/test_files/simplify_expr.slt

@@ -44,12 +44,27 @@ query TT
 explain select b from t where b !~ '.*'
 ----
 logical_plan
-01)Filter: t.b = Utf8View("")
+01)Filter: t.b IS NULL AND Boolean(NULL)
 02)--TableScan: t projection=[b]
 physical_plan
-01)FilterExec: b@0 =
+01)FilterExec: b@0 IS NULL AND NULL
 02)--DataSourceExec: partitions=1, partition_sizes=[1]
 
+query TB
+WITH vals(id, col) AS (
+    VALUES
+      (1, 'foo'::text),
+      (2, ''::text),
+      (3, NULL::text)
+)
+SELECT col, col !~ '.*'
+FROM vals
+ORDER BY id
+----
+foo false
+(empty) false
+NULL NULL
+
 query T
 select b from t where b ~ '.*'
 ----