drill-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From paul-rogers <...@git.apache.org>
Subject [GitHub] drill pull request #822: DRILL-5457: Spill implementation for Hash Aggregate
Date Sat, 27 May 2017 05:38:06 GMT
Github user paul-rogers commented on a diff in the pull request:

    --- Diff: exec/java-exec/src/main/java/org/apache/drill/exec/physical/impl/aggregate/HashAggTemplate.java
    @@ -230,15 +452,35 @@ public void setup(HashAggregate hashAggrConfig, HashTableConfig
htConfig, Fragme
           throw new IllegalArgumentException("Wrong number of workspace variables.");
    -//    this.context = context;
    +    this.context = context;
         this.stats = stats;
    -    this.allocator = allocator;
    +    this.allocator = oContext.getAllocator();
    +    this.oContext = oContext;
         this.incoming = incoming;
    -//    this.schema = incoming.getSchema();
         this.outgoing = outgoing;
         this.outContainer = outContainer;
    +    this.operatorId = hashAggrConfig.getOperatorId();
    +    is2ndPhase = hashAggrConfig.getAggPhase() == AggPrelBase.OperatorPhase.PHASE_2of2;
    +    isTwoPhase = hashAggrConfig.getAggPhase() != AggPrelBase.OperatorPhase.PHASE_1of1;
    +    canSpill = isTwoPhase; // single phase can not spill
    --- End diff --
    Here we have three related booleans, or 2^8 cases. Consider using an enum to identify
the (likely much smaller) number of actual cases. Maybe `ONE_PASS, FIRST_PHASE, SECOND_PHASE`?
    Then if the code does lots of "if this phase do that" kind of logic, it may be handy to
have a single base class with common logic, then three (or whatever) base classes that define
the phase-specific logic. Much easier to test and understand.

If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.

View raw message