From 7e83932f13238dc399651a87468d3eee7d444b2d Mon Sep 17 00:00:00 2001
From: Weizhen Wang <wangweizhen@pingcap.com>
Date: Fri, 5 Jan 2024 16:17:04 +0800
Subject: [PATCH] statistics: avoid allocating the memory when to auto analyze
 with pseudo table (#50099)

close pingcap/tidb#50100
---
 .../handle/autoanalyze/autoanalyze.go         |  6 ++---
 pkg/statistics/handle/handle.go               | 22 ++++++++++++++-----
 pkg/statistics/handle/types/interfaces.go     |  3 +++
 3 files changed, 23 insertions(+), 8 deletions(-)

diff --git a/pkg/statistics/handle/autoanalyze/autoanalyze.go b/pkg/statistics/handle/autoanalyze/autoanalyze.go
index 9e1fd89ed956e..f59e90d28fdbd 100644
--- a/pkg/statistics/handle/autoanalyze/autoanalyze.go
+++ b/pkg/statistics/handle/autoanalyze/autoanalyze.go
@@ -414,7 +414,7 @@ func RandomPickOneTableAndTryAutoAnalyze(
 			pi := tblInfo.GetPartitionInfo()
 			// No partitions, analyze the whole table.
 			if pi == nil {
-				statsTbl := statsHandle.GetTableStats(tblInfo)
+				statsTbl := statsHandle.GetTableStatsForAutoAnalyze(tblInfo)
 				sql := "analyze table %n.%n"
 				analyzed := tryAutoAnalyzeTable(sctx, statsHandle, sysProcTracker, tblInfo, statsTbl, autoAnalyzeRatio, sql, db, tblInfo.Name.O)
 				if analyzed {
@@ -491,10 +491,10 @@ func tryAutoAnalyzeTable(
 	sql string,
 	params ...interface{},
 ) bool {
-	// 1. If the stats are not loaded, we don't need to analyze it.
+	// 1. If the statistics are either not loaded or are classified as pseudo, there is no need for analyze
 	// 2. If the table is too small, we don't want to waste time to analyze it.
 	//    Leave the opportunity to other bigger tables.
-	if statsTbl.Pseudo || statsTbl.RealtimeCount < AutoAnalyzeMinCnt {
+	if statsTbl == nil || statsTbl.RealtimeCount < AutoAnalyzeMinCnt || statsTbl.Pseudo {
 		return false
 	}
 
diff --git a/pkg/statistics/handle/handle.go b/pkg/statistics/handle/handle.go
index 82872c1ebb72e..dd523bb8a7a8a 100644
--- a/pkg/statistics/handle/handle.go
+++ b/pkg/statistics/handle/handle.go
@@ -149,9 +149,18 @@ func (h *Handle) GetTableStats(tblInfo *model.TableInfo) *statistics.Table {
 	return h.GetPartitionStats(tblInfo, tblInfo.ID)
 }
 
+// GetTableStatsForAutoAnalyze is to get table stats but it will
+func (h *Handle) GetTableStatsForAutoAnalyze(tblInfo *model.TableInfo) *statistics.Table {
+	return h.getPartitionStats(tblInfo, tblInfo.ID, false)
+}
+
 // GetPartitionStats retrieves the partition stats from cache.
 // TODO: remove GetTableStats later on.
 func (h *Handle) GetPartitionStats(tblInfo *model.TableInfo, pid int64) *statistics.Table {
+	return h.getPartitionStats(tblInfo, pid, true)
+}
+
+func (h *Handle) getPartitionStats(tblInfo *model.TableInfo, pid int64, returnPseudo bool) *statistics.Table {
 	var tbl *statistics.Table
 	if h == nil {
 		tbl = statistics.PseudoTable(tblInfo, false)
@@ -160,12 +169,15 @@ func (h *Handle) GetPartitionStats(tblInfo *model.TableInfo, pid int64) *statist
 	}
 	tbl, ok := h.Get(pid)
 	if !ok {
-		tbl = statistics.PseudoTable(tblInfo, false)
-		tbl.PhysicalID = pid
-		if tblInfo.GetPartitionInfo() == nil || h.Len() < 64 {
-			h.UpdateStatsCache([]*statistics.Table{tbl}, nil)
+		if returnPseudo {
+			tbl = statistics.PseudoTable(tblInfo, false)
+			tbl.PhysicalID = pid
+			if tblInfo.GetPartitionInfo() == nil || h.Len() < 64 {
+				h.UpdateStatsCache([]*statistics.Table{tbl}, nil)
+			}
+			return tbl
 		}
-		return tbl
+		return nil
 	}
 	return tbl
 }
diff --git a/pkg/statistics/handle/types/interfaces.go b/pkg/statistics/handle/types/interfaces.go
index d8eb1c7663ac7..b7b7a65fd3130 100644
--- a/pkg/statistics/handle/types/interfaces.go
+++ b/pkg/statistics/handle/types/interfaces.go
@@ -430,6 +430,9 @@ type StatsHandle interface {
 	// GetTableStats retrieves the statistics table from cache, and the cache will be updated by a goroutine.
 	GetTableStats(tblInfo *model.TableInfo) *statistics.Table
 
+	// GetTableStatsForAutoAnalyze retrieves the statistics table from cache, but it will not return pseudo.
+	GetTableStatsForAutoAnalyze(tblInfo *model.TableInfo) *statistics.Table
+
 	// GetPartitionStats retrieves the partition stats from cache.
 	GetPartitionStats(tblInfo *model.TableInfo, pid int64) *statistics.Table