Change in asterixdb[master]: [ASTERIXDB-3208][COMP] Fix for array predicate selectivity

[AsterixDB Code Review]

>From Vijay Sarathy <vijay.sara...@couchbase.com>:

Vijay Sarathy has uploaded this change for review. ( 
https://asterix-gerrit.ics.uci.edu/c/asterixdb/+/17626 )


Change subject: [ASTERIXDB-3208][COMP] Fix for array predicate selectivity
......................................................................

[ASTERIXDB-3208][COMP] Fix for array predicate selectivity

Change-Id: I890b5c2a32b583a8d6e1f23c5f27d2c912ce3ef9
---
M 
asterixdb/asterix-algebra/src/main/java/org/apache/asterix/optimizer/rules/cbo/Stats.java
M 
asterixdb/asterix-algebra/src/main/java/org/apache/asterix/optimizer/rules/cbo/JoinNode.java
2 files changed, 36 insertions(+), 4 deletions(-)



  git pull ssh://asterix-gerrit.ics.uci.edu:29418/asterixdb 
refs/changes/26/17626/1

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 7f0a749..6c5b2ca 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 @@
                         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 785d56b..b285de2 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 @@
     }

     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 @@
                 }
             }
         }
-        // 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;
     }

--
To view, visit https://asterix-gerrit.ics.uci.edu/c/asterixdb/+/17626
To unsubscribe, or for help writing mail filters, visit 
https://asterix-gerrit.ics.uci.edu/settings

Gerrit-Project: asterixdb
Gerrit-Branch: master
Gerrit-Change-Id: I890b5c2a32b583a8d6e1f23c5f27d2c912ce3ef9
Gerrit-Change-Number: 17626
Gerrit-PatchSet: 1
Gerrit-Owner: Vijay Sarathy <vijay.sara...@couchbase.com>
Gerrit-MessageType: newchange