[asterixdb] branch master updated: [ASTERIXDB-3208][COMP] Fix for array predicate selectivity

[vsarathy1]

This is an automated email from the ASF dual-hosted git repository.

vsarathy1 pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/asterixdb.git


The following commit(s) were added to refs/heads/master by this push:
     new 290d5374a8 [ASTERIXDB-3208][COMP] Fix for array predicate selectivity
290d5374a8 is described below

commit 290d5374a80b9e62bfe01f87ed689a0513d603e7
Author: Vijay Sarathy <vijay.sara...@couchbase.com>
AuthorDate: Fri Jun 30 12:38:51 2023 -0700

    [ASTERIXDB-3208][COMP] Fix for array predicate selectivity
    
    Change-Id: I890b5c2a32b583a8d6e1f23c5f27d2c912ce3ef9
    Reviewed-on: https://asterix-gerrit.ics.uci.edu/c/asterixdb/+/17626
    Integration-Tests: Jenkins <jenk...@fulliautomatix.ics.uci.edu>
    Reviewed-by: Vijay Sarathy <vijay.sara...@couchbase.com>
    Reviewed-by: Ali Alsuliman <ali.al.solai...@gmail.com>
    Tested-by: Jenkins <jenk...@fulliautomatix.ics.uci.edu>
---
 .../asterix/optimizer/rules/cbo/JoinNode.java      |  3 ++-
 .../apache/asterix/optimizer/rules/cbo/Stats.java  | 28 +++++++++++++++++++---
 2 files changed, 27 insertions(+), 4 deletions(-)

diff --git 
a/asterixdb/asterix-algebra/src/main/java/org/apache/asterix/optimizer/rules/cbo/JoinNode.java
 
b/asterixdb/asterix-algebra/src/main/java/org/apache/asterix/optimizer/rules/cbo/JoinNode.java
index 7f0a74932a..6c5b2ca12b 100644
--- 
a/asterixdb/asterix-algebra/src/main/java/org/apache/asterix/optimizer/rules/cbo/JoinNode.java
+++ 
b/asterixdb/asterix-algebra/src/main/java/org/apache/asterix/optimizer/rules/cbo/JoinNode.java
@@ -471,7 +471,8 @@ public class JoinNode {
                         selOp = new SelectOperator(new MutableObject<>(afce));
                         selOp.getInputs().add(new MutableObject<>(leafInput));
                     }
-                    sel = 
joinEnum.getStatsHandle().findSelectivityForThisPredicate(selOp, afce, 
this.origCardinality);
+                    sel = 
joinEnum.getStatsHandle().findSelectivityForThisPredicate(selOp, afce,
+                            
chosenIndex.getIndexType().equals(DatasetConfig.IndexType.ARRAY), 
this.origCardinality);
                 }
                 IndexCostInfo.add(new Triple<>(chosenIndex, sel, afce));
             }
diff --git 
a/asterixdb/asterix-algebra/src/main/java/org/apache/asterix/optimizer/rules/cbo/Stats.java
 
b/asterixdb/asterix-algebra/src/main/java/org/apache/asterix/optimizer/rules/cbo/Stats.java
index 785d56b49c..b285de280b 100644
--- 
a/asterixdb/asterix-algebra/src/main/java/org/apache/asterix/optimizer/rules/cbo/Stats.java
+++ 
b/asterixdb/asterix-algebra/src/main/java/org/apache/asterix/optimizer/rules/cbo/Stats.java
@@ -372,7 +372,7 @@ public class Stats {
     }
 
     protected double findSelectivityForThisPredicate(SelectOperator selOp, 
AbstractFunctionCallExpression exp,
-            double datasetCard) throws AlgebricksException {
+            boolean arrayIndex, double datasetCard) throws AlgebricksException 
{
         // replace the SelOp.condition with the new exp and replace it at the 
end
         // The Selop here is the start of the leafInput.
 
@@ -442,13 +442,35 @@ public class Stats {
                 }
             }
         }
-        // switch  the scanOp back
-        parent.getInputs().get(0).setValue(scanOp);
 
         double predicateCardinality = (double) ((AInt64) 
result.get(0).get(0)).getLongValue();
         if (predicateCardinality == 0.0) {
             predicateCardinality = 0.0001 * 
idxDetails.getSampleCardinalityTarget();
         }
+
+        if (arrayIndex) {
+            // In case of array predicates, the sample cardinality should be 
computed as
+            // the number of unnested array elements. Run a second sampling 
query to compute this.
+            // The query should already have the unnest operation, so simply 
replace the select clause with TRUE
+            // to get the unnested cardinality from the sample.
+            // Example query: SELECT count(*) as revenue
+            //                FROM   orders o, o.o_orderline ol
+            //                WHERE  ol.ol_delivery_d  >= '2016-01-01 
00:00:00.000000'
+            //                  AND  ol.ol_delivery_d < '2017-01-01 
00:00:00.000000';
+            // ol_delivery_d is part of the array o_orderline
+            // To get the unnested cardinality,we run the following query on 
the sample:
+            // SELECT count(*) as revenue
+            // FROM   orders o, o.o_orderline ol
+            // WHERE  TRUE;
+            ILogicalExpression saveExprs = selOp.getCondition().getValue();
+            selOp.getCondition().setValue(ConstantExpression.TRUE);
+            result = runSamplingQuery(optCtx, selOp);
+            selOp.getCondition().setValue(saveExprs);
+            sampleCard = (double) ((AInt64) 
result.get(0).get(0)).getLongValue();
+        }
+        // switch  the scanOp back
+        parent.getInputs().get(0).setValue(scanOp);
+
         double sel = (double) predicateCardinality / sampleCard;
         return sel;
     }