statsitstics: avoid sync load column which is skiped type to analyze (#57144) (#57615)

close #57138

Author: ti-chi-bot

pkg/statistics/handle/syncload/BUILD.bazel

@@ -14,6 +14,7 @@ go_library(
         "//pkg/parser/mysql",
         "//pkg/sessionctx",
         "//pkg/sessionctx/stmtctx",
+        "//pkg/sessionctx/variable",
         "//pkg/statistics",
         "//pkg/statistics/handle/storage",
         "//pkg/statistics/handle/types",
@@ -34,7 +35,7 @@ go_test(
     srcs = ["stats_syncload_test.go"],
     flaky = True,
     race = "on",
-    shard_count = 7,
+    shard_count = 8,
     deps = [
         ":syncload",
         "//pkg/config",

pkg/statistics/handle/syncload/stats_syncload.go

@@ -30,6 +30,7 @@ import (
 	"github.com/pingcap/tidb/pkg/parser/mysql"
 	"github.com/pingcap/tidb/pkg/sessionctx"
 	"github.com/pingcap/tidb/pkg/sessionctx/stmtctx"
+	"github.com/pingcap/tidb/pkg/sessionctx/variable"
 	"github.com/pingcap/tidb/pkg/statistics"
 	"github.com/pingcap/tidb/pkg/statistics/handle/storage"
 	statstypes "github.com/pingcap/tidb/pkg/statistics/handle/types"
@@ -306,6 +307,14 @@ func (s *statsSyncLoad) handleOneItemTask(task *statstypes.NeededItemTask) (err
 			s.statsHandle.SPool().Destroy(se)
 		}
 	}()
+	var skipTypes map[string]struct{}
+	val, err := sctx.GetSessionVars().GlobalVarsAccessor.GetGlobalSysVar(variable.TiDBAnalyzeSkipColumnTypes)
+	if err != nil {
+		logutil.BgLogger().Warn("failed to get global variable", zap.Error(err))
+	} else {
+		skipTypes = variable.ParseAnalyzeSkipColumnTypes(val)
+	}
+
 	item := task.Item.TableItemID
 	statsTbl, ok := s.statsHandle.Get(item.TableID)
 
@@ -342,6 +351,13 @@ func (s *statsSyncLoad) handleOneItemTask(task *statstypes.NeededItemTask) (err
 			// so we have to get the column info from the domain.
 			wrapper.colInfo = tblInfo.GetColumnByID(item.ID)
 		}
+		if skipTypes != nil {
+			_, skip := skipTypes[types.TypeToStr(wrapper.colInfo.FieldType.GetType(), wrapper.colInfo.FieldType.GetCharset())]
+			if skip {
+				return nil
+			}
+		}
+
 		// If this column is not analyzed yet and we don't have it in memory.
 		// We create a fake one for the pseudo estimation.
 		// Otherwise, it will trigger the sync/async load again, even if the column has not been analyzed.
@@ -351,6 +367,7 @@ func (s *statsSyncLoad) handleOneItemTask(task *statstypes.NeededItemTask) (err
 			return nil
 		}
 	}
+	failpoint.Inject("handleOneItemTaskPanic", nil)
 	t := time.Now()
 	needUpdate := false
 	wrapper, err = s.readStatsForOneItem(sctx, item, wrapper, isPkIsHandle, task.Item.FullLoad)

pkg/statistics/handle/syncload/stats_syncload_test.go

@@ -56,6 +56,22 @@ func TestSyncLoadSkipUnAnalyzedItems(t *testing.T) {
 	failpoint.Disable("github.com/pingcap/tidb/pkg/statistics/handle/syncload/assertSyncLoadItems")
 }
 
+func TestSyncLoadSkipAnalyzSkipColumnItems(t *testing.T) {
+	store, dom := testkit.CreateMockStoreAndDomain(t)
+	tk := testkit.NewTestKit(t, store)
+	tk.MustExec("use test")
+	tk.MustExec("drop table if exists t")
+	tk.MustExec("create table t(`id` bigint(20) NOT NULL AUTO_INCREMENT,content text,PRIMARY KEY (`id`))")
+	h := dom.StatsHandle()
+	h.SetLease(1)
+
+	tk.MustExec("analyze table t")
+	tk.MustExec("set @@session.tidb_analyze_skip_column_types = 'json, text, blob'") // text is not default.
+	require.NoError(t, failpoint.Enable("github.com/pingcap/tidb/pkg/statistics/handle/syncload/handleOneItemTaskPanic", `panic`))
+	tk.MustQuery("trace plan select * from t where content ='ab'")
+	require.NoError(t, failpoint.Disable("github.com/pingcap/tidb/pkg/statistics/handle/syncload/handleOneItemTaskPanic"))
+}
+
 func TestConcurrentLoadHist(t *testing.T) {
 	store, dom := testkit.CreateMockStoreAndDomain(t)