modify cbo statistics in filter and scan

yx-keith · May 22, 2024 · 487524a · 487524a
1 parent bb407ee
commit 487524a
Show file tree

Hide file tree

Showing 3 changed files with 25 additions and 3 deletions.
diff --git a/fe/fe-core/src/main/java/org/apache/doris/nereids/stats/FilterEstimation.java b/fe/fe-core/src/main/java/org/apache/doris/nereids/stats/FilterEstimation.java
@@ -18,6 +18,7 @@
 package org.apache.doris.nereids.stats;
 
 import org.apache.doris.analysis.LiteralExpr;
+import org.apache.doris.nereids.CascadesContext;
 import org.apache.doris.nereids.stats.FilterEstimation.EstimationContext;
 import org.apache.doris.nereids.trees.TreeNode;
 import org.apache.doris.nereids.trees.expressions.And;
@@ -92,6 +93,20 @@ public Statistics estimate(Expression expression, Statistics statistics) {
         return stats;
     }
 
+    /**
+     * This method will update the stats according to the selectivity.
+     */
+    public Statistics estimate(Expression expression, Statistics statistics, CascadesContext context) {
+        // For a comparison predicate, only when it's left side is a slot and right side is a literal, we would
+        // consider is a valid predicate.
+        Statistics stats = expression.accept(this, new EstimationContext(statistics));
+        if (context.getConnectContext().getSessionVariable().enableModifyCboStatistics && stats.getRowCount() <= 0) {
+            stats = new StatisticsBuilder(stats).setRowCount(1).build();
+        }
+        stats.enforceValid();
+        return stats;
+    }
+
     @Override
     public Statistics visit(Expression expr, EstimationContext context) {
         return context.statistics.withSel(DEFAULT_INEQUALITY_COEFFICIENT);

diff --git a/fe/fe-core/src/main/java/org/apache/doris/nereids/stats/StatsCalculator.java b/fe/fe-core/src/main/java/org/apache/doris/nereids/stats/StatsCalculator.java
@@ -614,7 +614,7 @@ private Statistics computeFilter(Filter filter) {
                         .map(NamedExpression::toSlot).collect(Collectors.toSet());
                 Expression predicate = filter.getPredicate();
                 if (predicate.anyMatch(s -> slots.contains(s))) {
-                    return new FilterEstimation(slots).estimate(filter.getPredicate(), stats);
+                    return new FilterEstimation(slots).estimate(filter.getPredicate(), stats, this.cascadesContext);
                 }
             } else if (plan instanceof LogicalJoin && filter instanceof LogicalFilter
                     && filter.getConjuncts().stream().anyMatch(e -> e instanceof IsNull)) {
@@ -636,8 +636,7 @@ private Statistics computeFilter(Filter filter) {
                 }
             }
         }
-
-        return new FilterEstimation(isOnBaseTable).estimate(filter.getPredicate(), stats);
+        return new FilterEstimation().estimate(filter.getPredicate(), stats, this.cascadesContext);
     }
 
     private Statistics computeGeneratedIsNullStats(LogicalJoin join, Filter filter) {
@@ -771,6 +770,9 @@ private Statistics computeCatalogRelation(CatalogRelation catalogRelation) {
         TableStatsMeta tableMeta = analysisManager.findTableStatsStatus(table.getId());
         long tableUpdatedRows = tableMeta == null ? 0 : tableMeta.updatedRows.get();
         double rowCount = catalogRelation.getTable().getRowCountForNereids();
+        if (this.cascadesContext.getConnectContext().getSessionVariable().enableModifyCboStatistics && rowCount <= 0) {
+            rowCount = 1.0;
+        }
         boolean hasUnknownKeyCol = false;
         long idxId = -1;
         if (catalogRelation instanceof OlapScan) {

diff --git a/fe/fe-core/src/main/java/org/apache/doris/qe/SessionVariable.java b/fe/fe-core/src/main/java/org/apache/doris/qe/SessionVariable.java
@@ -349,6 +349,8 @@ public class SessionVariable implements Serializable, Writable {
 
     public static final String ENABLE_CBO_STATISTICS = "enable_cbo_statistics";
 
+    public static final String ENABLE_MODIFY_CBO_STATISTICS = "enable_modify_cbo_statistics";
+
     public static final String ENABLE_SAVE_STATISTICS_SYNC_JOB = "enable_save_statistics_sync_job";
 
     public static final String ENABLE_ELIMINATE_SORT_NODE = "enable_eliminate_sort_node";
@@ -1316,6 +1318,9 @@ public void setEnableLeftZigZag(boolean enableLeftZigZag) {
     @VariableMgr.VarAttr(name = ENABLE_CBO_STATISTICS)
     public boolean enableCboStatistics = false;
 
+    @VariableMgr.VarAttr(name = ENABLE_MODIFY_CBO_STATISTICS, needForward = true)
+    public boolean enableModifyCboStatistics = true;
+
     @VariableMgr.VarAttr(name = ENABLE_ELIMINATE_SORT_NODE)
     public boolean enableEliminateSortNode = true;