Merge pull request #56253 from ClickHouse/backport/23.8/54696

Backport #54696 to 23.8: Fix: avoid using regex match, possibly containing alternation, as a key condition.
ClickHouse · Nov 2, 2023 · 2a5df64 · 2a5df64
2 parents f6d503e + 378b062
commit 2a5df64
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 1 deletion.
diff --git a/src/Storages/MergeTree/KeyCondition.cpp b/src/Storages/MergeTree/KeyCondition.cpp
@@ -421,7 +421,12 @@ const KeyCondition::AtomMap KeyCondition::atom_map
             if (value.getType() != Field::Types::String)
                 return false;
 
-            String prefix = extractFixedPrefixFromRegularExpression(value.get<const String &>());
+            const String & expression = value.get<const String &>();
+            // This optimization can't process alternation - this would require a comprehensive parsing of regular expression.
+            if (expression.contains('|'))
+                return false;
+
+            String prefix = extractFixedPrefixFromRegularExpression(expression);
             if (prefix.empty())
                 return false;
 

diff --git a/tests/queries/0_stateless/02462_match_regexp_pk.sql b/tests/queries/0_stateless/02462_match_regexp_pk.sql
@@ -7,3 +7,4 @@ SELECT count() FROM mt_match_pk WHERE match(v, '^ab');
 SELECT count() FROM mt_match_pk WHERE match(v, '^a.');
 SELECT count() FROM mt_match_pk WHERE match(v, '^ab*');
 SELECT count() FROM mt_match_pk WHERE match(v, '^ac?');
+SELECT count() FROM mt_match_pk WHERE match(v, '^a$|^b'); -- {serverError INDEX_NOT_USED}