[
https://issues.apache.org/jira/browse/MADLIB-973?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Frank McQuillan closed MADLIB-973.
----------------------------------
> Path - make aggregate computation per pattern match (not per partition)
> -----------------------------------------------------------------------
>
> Key: MADLIB-973
> URL: https://issues.apache.org/jira/browse/MADLIB-973
> Project: Apache MADlib
> Issue Type: New Feature
> Components: Module: Utilities
> Reporter: Frank McQuillan
> Assignee: Rahul Iyer
> Fix For: v1.9
>
> Attachments: Ecommerce data set for path test 3.csv, path query3.sql
>
>
> Story:
> As a user, I want to be able to define aggregates on pattern matches, so that
> I can derive useful information from the patterns that I am interested in.
> In addition, I want to be able to do richer subsequent queries on the pattern
> matches, so I want need match_id and symbol added to the output table of
> matched tuples.
> Other:
> 1) This a progression from these stores
> https://issues.apache.org/jira/browse/MADLIB-916
> https://issues.apache.org/jira/browse/MADLIB-917
> https://issues.apache.org/jira/browse/MADLIB-931
> which were doing aggregates on partitions as opposed to pattern matches.
> Acceptance test:
> Ecommerce data set for path test 3.csv
> Pattern match: IMPR->CLICK->CONV
> i.e., exactly 1 of each in succession
> notes:
> an IMPR is defined by no CLICK and no CONV
> all CONV events also have a CLICK event, but they are considered to be only
> CONV events
> pattern matches do not cross day boundaries
> “Path Match” column indicates a path that matches the above pattern
> output:
> a) partition by user ID
> total revenue and total margin by user ID
> 101121
> match ID = 1
> rev = 131
> margin = 28
> 101331
> match ID = 1
> rev = 112
> margin = 36
> 101331
> match ID = 2
> rev = 456
> margin = 77
> Note there are 3 output rows since there are 3 matches.
> Also, please output match ID, which is a count from 1->N within the
> partition, where N is the number of matches in that partition. It means
> partition ID + match ID is unique across the result set.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
