[
https://issues.apache.org/jira/browse/DRILL-35?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Timothy Chen updated DRILL-35:
------------------------------
Comment: was deleted
(was: A bit confused what this operator suppose to output. I thought the
existing AggregateSimpleROP is the impl of CollapsingAggregator, but it doesn't
define carryovers, etc.
So according to the plan doc, the list of aggregates are re-evaluated on each
record which I will interpret that it does the aggregation on a per record
basis, so given rows:
ppu: 15
sales: 10
ppu: 10
sales: 30
the aggregate of summary = sum(ppu * sales) will output
summary = 450
Which differs from CollapseAggregate where it has a key input for boundry, and
aggregations are done per boundry.
However, the within field can specify a boundry which I'm not sure what that
meant? Does this mean the aggregation is done within a boundry?)
> Implement RunningAggregate Reference Operator
> ---------------------------------------------
>
> Key: DRILL-35
> URL: https://issues.apache.org/jira/browse/DRILL-35
> Project: Apache Drill
> Issue Type: New Feature
> Reporter: Timothy Chen
> Assignee: Timothy Chen
> Labels: logical
>
> RunningAggregate (1)
> The running aggregate operator takes an input record and adds appends a set
> of running aggregations and outputs the resulting record. The aggregations
> are re-evaluated on each record within the incoming segment in the order they
> are provided. Segment focus can be defined with the ‘within’ value such that
> aggregations are reset at each segment boundary.
> { @id†: <opref>, op: “runningaggregate”,
> input†: <input>,
> within*: <name>,
> aggregations: [
> {ref: <name>, expr: <aggexpr> },...
> ]
> }
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
