From d06fd67764e6f3d21d73e5d658677a12d65c8dfd Mon Sep 17 00:00:00 2001
From: Zhou Kunqin <25057648+time-and-fate@users.noreply.github.com>
Date: Tue, 17 Aug 2021 16:02:00 +0800
Subject: [PATCH] statistics: fix "data too long" error when dumping stats from
 table with new collation data (#27033)

---
 executor/analyze.go              |  2 +-
 statistics/handle/handle.go      |  6 ++++++
 statistics/handle/handle_test.go | 21 +++++++++++++++++++++
 3 files changed, 28 insertions(+), 1 deletion(-)

diff --git a/executor/analyze.go b/executor/analyze.go
index 67cc4cbd54..0f3bcee345 100644
--- a/executor/analyze.go
+++ b/executor/analyze.go
@@ -1270,7 +1270,7 @@ workLoop:
 				// When it's new collation data, we need to use its collate key instead of original value because only
 				// the collate key can ensure the correct ordering.
 				// This is also corresponding to similar operation in (*statistics.Column).GetColumnRowCount().
-				if ft.EvalType() == types.ETString {
+				if ft.EvalType() == types.ETString && ft.Tp != mysql.TypeEnum && ft.Tp != mysql.TypeSet {
 					val.SetBytes(collate.GetCollator(ft.Collate).Key(val.GetString()))
 				}
 				sampleItems = append(sampleItems, &statistics.SampleItem{
diff --git a/statistics/handle/handle.go b/statistics/handle/handle.go
index 48e4997c56..1c2a89fe0f 100644
--- a/statistics/handle/handle.go
+++ b/statistics/handle/handle.go
@@ -1287,6 +1287,12 @@ func (h *Handle) histogramFromStorage(reader *statsReader, tableID int64, colID
 		} else {
 			sc := &stmtctx.StatementContext{TimeZone: time.UTC}
 			d := rows[i].GetDatum(2, &fields[2].Column.FieldType)
+			// When there's new collation data, the length of bounds of histogram(the collate key) might be
+			// longer than the FieldType.Flen of this column.
+			// We change it to TypeBlob to bypass the length check here.
+			if tp.EvalType() == types.ETString && tp.Tp != mysql.TypeEnum && tp.Tp != mysql.TypeSet {
+				tp = types.NewFieldType(mysql.TypeBlob)
+			}
 			lowerBound, err = d.ConvertTo(sc, tp)
 			if err != nil {
 				return nil, errors.Trace(err)
diff --git a/statistics/handle/handle_test.go b/statistics/handle/handle_test.go
index 9cfd064477..533a15ad54 100644
--- a/statistics/handle/handle_test.go
+++ b/statistics/handle/handle_test.go
@@ -2867,6 +2867,27 @@ func (s *testSerialStatsSuite) TestCorrelationWithDefinedCollate(c *C) {
 	c.Assert(rows[0][5], Equals, "-1.000000")
 }
 
+func (s *testSerialStatsSuite) TestLoadHistogramWithCollate(c *C) {
+	defer cleanEnv(c, s.store, s.do)
+	testKit := testkit.NewTestKit(c, s.store)
+	collate.SetNewCollationEnabledForTest(true)
+	defer collate.SetNewCollationEnabledForTest(false)
+	testKit.MustExec("use test")
+	testKit.MustExec("drop table if exists t")
+	testKit.MustExec("create table t(a varchar(10) collate utf8mb4_unicode_ci);")
+	testKit.MustExec("insert into t values('abcdefghij');")
+	testKit.MustExec("insert into t values('abcdufghij');")
+	testKit.MustExec("analyze table t with 0 topn;")
+	do := s.do
+	h := do.StatsHandle()
+	is := do.InfoSchema()
+	tbl, err := is.TableByName(model.NewCIStr("test"), model.NewCIStr("t"))
+	c.Assert(err, IsNil)
+	tblInfo := tbl.Meta()
+	_, err = h.TableStatsFromStorage(tblInfo, tblInfo.ID, true, 0)
+	c.Assert(err, IsNil)
+}
+
 func (s *testSerialStatsSuite) TestFastAnalyzeColumnHistWithNullValue(c *C) {
 	defer cleanEnv(c, s.store, s.do)
 	testKit := testkit.NewTestKit(c, s.store)