expression: fix substring_index crash tiflash in some case (#9137) (#…

…9140)

close #9116

Fix tiflash crash caused by function `substring_index` in some corner cases

Co-authored-by: wshwsh12 <793703860@qq.com>
Co-authored-by: ti-chi-bot[bot] <108142056+ti-chi-bot[bot]@users.noreply.github.com>

dbms/src/Functions/FunctionsString.cpp

@@ -5268,6 +5268,7 @@ class FunctionSubStringIndex : public IFunction
  const UInt8 * pos = begin;
  const UInt8 * end = pos + data_size;
  assert(delim_size != 0);
+ assert(count != 0);
  if (count > 0)
  {
  // Fast exit when count * delim_size > data_size
@@ -5283,10 +5284,11 @@ class FunctionSubStringIndex : public IFunction
  if (match == end || count == 0)
  {
  copyDataToResult(res_data, res_offset, begin, match);
- break;
+ return;
  }
  pos = match + delim_size;
  }
+ copyDataToResult(res_data, res_offset, begin, end);
  }
  else
  {

dbms/src/Functions/tests/gtest_substring_index.cpp

@@ -364,6 +364,15 @@ try
  createColumn<Nullable<String>>({"www.pingcap.com", "www...www", "中文.测.试。。。", "www.www"}),
  createColumn<Nullable<String>>({"", "", "", ""}),
  createColumn<Nullable<Int64>>({2, 2, 2, 2})));
+
+ // Test issue 9116
+ ASSERT_COLUMN_EQ(
+ createColumn<Nullable<String>>({"aaabbba", "aaabbbaa", "aaabbbaaa", "aaabbbaaa", "aaabbbaaa"}),
+ executeFunction(
+ func_name,
+ createColumn<Nullable<String>>({"aaabbbaaa", "aaabbbaaa", "aaabbbaaa", "aaabbbaaa", "aaabbbaaa"}),
+ createColumn<Nullable<String>>({"a", "a", "a", "a", "a"}),
+ createColumn<Nullable<Int64>>({5, 6, 7, 8, 9})));
 }
 CATCH