statistics: fix a bug which causes panic when using the clustered ind…

…ex and the new collation (#21379)

Signed-off-by: wjhuang2016 <huangwenjun1997@gmail.com>

expression/integration_test.go

@@ -8029,3 +8029,22 @@ func (s *testIntegrationSuite) TestIssue19892(c *C) {
 		tk.MustQuery("SELECT c FROM dd").Check(testkit.Rows("0000-00-00 00:00:00"))
 	}
 }
+
+func (s *testIntegrationSerialSuite) TestIssue20876(c *C) {
+	collate.SetNewCollationEnabledForTest(true)
+	defer collate.SetNewCollationEnabledForTest(false)
+	tk := testkit.NewTestKit(c, s.store)
+	tk.MustExec("set @@tidb_enable_clustered_index=1;")
+	tk.MustExec("use test")
+	tk.MustExec("drop table if exists t;")
+	tk.MustExec("CREATE TABLE `t` (" +
+		"  `a` char(10) COLLATE utf8mb4_unicode_ci NOT NULL," +
+		"  `b` char(20) COLLATE utf8mb4_general_ci NOT NULL," +
+		"  `c` int(11) NOT NULL," +
+		"  PRIMARY KEY (`a`,`b`,`c`)," +
+		"  KEY `idx` (`a`)" +
+		")")
+	tk.MustExec("insert into t values ('#', 'C', 10), ('$', 'c', 20), ('$', 'c', 30), ('a', 'a', 10), ('A', 'A', 30)")
+	tk.MustExec("analyze table t")
+	tk.MustQuery("select * from t where a='#';").Check(testkit.Rows("# C 10"))
+}

statistics/histogram.go

@@ -820,18 +820,10 @@ func (c *Column) GetColumnRowCount(sc *stmtctx.StatementContext, ranges []*range
 		highVal := *rg.HighVal[0].Clone()
 		lowVal := *rg.LowVal[0].Clone()
 		if highVal.Kind() == types.KindString {
-			highVal.SetBytesAsString(collate.GetCollator(
-				highVal.Collation()).Key(highVal.GetString()),
-				highVal.Collation(),
-				uint32(highVal.Length()),
-			)
+			highVal.SetBytes(collate.GetCollator(highVal.Collation()).Key(highVal.GetString()))
 		}
 		if lowVal.Kind() == types.KindString {
-			lowVal.SetBytesAsString(collate.GetCollator(
-				lowVal.Collation()).Key(lowVal.GetString()),
-				lowVal.Collation(),
-				uint32(lowVal.Length()),
-			)
+			lowVal.SetBytes(collate.GetCollator(lowVal.Collation()).Key(lowVal.GetString()))
 		}
 		cmp, err := lowVal.CompareDatum(sc, &highVal)
 		if err != nil {