[BugFix] Keep non-null rows for negated empty-literal MATCH in inverted-index seek

seek_inverted_index cleared the row bitmap whenever the match literal was empty
(read_inverted_index returned nullopt), regardless of negation. That dropped
every row for NOT (col MATCH ''), even though MATCH '' matches nothing and its
negation is TRUE for all non-null rows. This surfaced once the OR-MATCH fallback
started reusing seek_inverted_index.

Substitute an empty posting set for the nullopt case and reuse the shared
with_not / intersect path, so col MATCH '' keeps no rows while NOT (col MATCH '')
keeps every non-null row (only NULLs stay out). Fixes the same latent bug on the
AND filtering path. Add SQL regression for both the AND and OR-MATCH fallback
paths.

Signed-off-by: eason <1569163626@qq.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>

Author: Beauuu-0826

be/src/storage_primitive/column_expr_predicate.cpp

@@ -390,10 +390,11 @@ Status ColumnExprPredicate::seek_inverted_index(const std::string& column_name,
 
     ASSIGN_OR_RETURN(auto roaring_opt, read_inverted_index(column_name, iterator));
 
-    // nullopt means empty match string - clear bitmap
+    // nullopt means an empty match string, which matches no rows. Fall through
+    // with an empty posting set so the shared logic below produces the right
+    // result
     if (!roaring_opt.has_value()) {
-        *row_bitmap -= *row_bitmap;
-        return Status::OK();
+        roaring_opt.emplace();
     }
 
     if (with_not) {

test/sql/test_inverted_index/R/test_inverted_index

@@ -778,6 +778,20 @@ SELECT * FROM t_gin_index_multiple_predicate_none WHERE text_column NOT match_an
 4	This is Gin Index
 5	None
 -- !result
+SELECT * FROM t_gin_index_multiple_predicate_none WHERE text_column NOT match_any "";
+-- result:
+1	ABC
+2	abc
+3	ABD
+4	This is Gin Index
+-- !result
+SELECT * FROM t_gin_index_multiple_predicate_none WHERE text_column NOT match_any "" OR id1 = 100;
+-- result:
+1	ABC
+2	abc
+3	ABD
+4	This is Gin Index
+-- !result
 DROP TABLE t_gin_index_multiple_predicate_none;
 -- result:
 -- !result

test/sql/test_inverted_index/T/test_inverted_index

@@ -320,6 +320,11 @@ SELECT * FROM t_gin_index_multiple_predicate_none WHERE text_column NOT match_an
 SELECT * FROM t_gin_index_multiple_predicate_none WHERE id1 = 100 OR text_column NOT match_any "ABC";
 -- A sibling OR arm may still legitimately select the NULL row.
 SELECT * FROM t_gin_index_multiple_predicate_none WHERE text_column NOT match_any "ABC" OR id1 = 5;
+-- Empty match literal: match_any '' matches nothing, so NOT is TRUE for every
+-- non-null row. Both the AND path and the OR-MATCH fallback must keep those
+-- rows (only the NULL row stays out), not drop them.
+SELECT * FROM t_gin_index_multiple_predicate_none WHERE text_column NOT match_any "";
+SELECT * FROM t_gin_index_multiple_predicate_none WHERE text_column NOT match_any "" OR id1 = 100;
 
 DROP TABLE t_gin_index_multiple_predicate_none;