Merge pull request #19393 from ClickHouse/extract-all-groups-fail-fast

Fail fast in incorrect usage of extractAllGroups

src/Functions/extractAllGroups.h

@@ -14,12 +14,14 @@
 
 #include <Core/iostream_debug_helpers.h>
 
+
 namespace DB
 {
 
 namespace ErrorCodes
 {
     extern const int BAD_ARGUMENTS;
+    extern const int TOO_LARGE_ARRAY_SIZE;
 }
 
 
@@ -145,11 +147,11 @@ public:
         }
         else
         {
-            std::vector<StringPiece> all_matches;
-            // number of times RE matched on each row of haystack column.
-            std::vector<size_t> number_of_matches_per_row;
+            PODArray<StringPiece, 0> all_matches;
+            /// Number of times RE matched on each row of haystack column.
+            PODArray<size_t, 0> number_of_matches_per_row;
 
-            // we expect RE to match multiple times on each row, `* 8` is arbitrary to reduce number of re-allocations.
+            /// We expect RE to match multiple times on each row, `* 8` is arbitrary to reduce number of re-allocations.
             all_matches.reserve(input_rows_count * groups_count * 8);
             number_of_matches_per_row.reserve(input_rows_count);
 
@@ -170,6 +172,12 @@ public:
                     for (size_t group = 1; group <= groups_count; ++group)
                         all_matches.push_back(matched_groups[group]);
 
+                    /// Additional limit to fail fast on supposedly incorrect usage.
+                    static constexpr size_t MAX_GROUPS_PER_ROW = 1000000;
+
+                    if (all_matches.size() > MAX_GROUPS_PER_ROW)
+                        throw Exception(ErrorCodes::TOO_LARGE_ARRAY_SIZE, "Too large array size in the result of function {}", getName());
+
                     pos = matched_groups[0].data() + std::max<size_t>(1, matched_groups[0].size());
 
                     ++matches_per_row;

tests/queries/0_stateless/01661_extract_all_groups_fail_fast.sql

@@ -0,0 +1 @@
+SELECT repeat('abcdefghijklmnopqrstuvwxyz', number * 100) AS haystack, extractAllGroupsHorizontal(haystack, '(\\w)') AS matches FROM numbers(1023); -- { serverError 128 }