[
https://issues.apache.org/jira/browse/MAHOUT-1615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14133557#comment-14133557
]
ASF GitHub Bot commented on MAHOUT-1615:
----------------------------------------
Github user dlyubimov commented on the pull request:
https://github.com/apache/mahout/pull/52#issuecomment-55551176
Oops. the script has error. sorry i do see some duplicates in drm C
val inCoreC = drmC.mapBlock() { case (keys,block) =>
keys.map(println)
keys -> block
}.collect
key-2
key-2
key-2
key-17
key-17
key-17
key-29
key-29
key-29
key-5
key-5
key-5
key-8
key-8
key-8
key-20
key-20
key-20
key-23
key-23
key-23
key-14
key-14
key-14
key-11
key-11
key-11
key-26
key-26
key-26
I did check row bindings on drmB and they are correct, but after
save-reload cycle they are no more correct. Which means collect() is not the
reason, it is either save or drmFromHdfs. I still think the patch is kludgy and
there has to be a simpler way to fix this though. Let me consider this for a
moment.
> SparkEngine drmFromHDFS returning the same Key for all Key,Vec Pairs for
> Text-Keyed SequenceFiles
> -------------------------------------------------------------------------------------------------
>
> Key: MAHOUT-1615
> URL: https://issues.apache.org/jira/browse/MAHOUT-1615
> Project: Mahout
> Issue Type: Bug
> Reporter: Andrew Palumbo
> Fix For: 1.0
>
>
> When reading in seq2sparse output from HDFS in the spark-shell of form
> <Text,VectorWriteable> SparkEngine's drmFromHDFS method is creating rdds
> with the same Key for all Pairs:
> {code}
> mahout> val drmTFIDF= drmFromHDFS( path =
> "/tmp/mahout-work-andy/20news-test-vectors/part-r-00000")
> {code}
> Has keys:
> {...}
> key: /talk.religion.misc/84570
> key: /talk.religion.misc/84570
> key: /talk.religion.misc/84570
> {...}
> for the entire set. This is the last Key in the set.
> The problem can be traced to the first line of drmFromHDFS(...) in
> SparkEngine.scala:
> {code}
> val rdd = sc.sequenceFile(path, classOf[Writable], classOf[VectorWritable],
> minPartitions = parMin)
> // Get rid of VectorWritable
> .map(t => (t._1, t._2.get()))
> {code}
> which gives the same key for all t._1.
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
