[
https://issues.apache.org/jira/browse/MAHOUT-1615?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14133249#comment-14133249
]
Andrew Palumbo edited comment on MAHOUT-1615 at 9/14/14 9:33 PM:
-----------------------------------------------------------------
In the mahout spark-shell I can get several other Keys (though not all):
{code}
mahout> val rdd = sdc.sequenceFile(path =
"/tmp/mahout-work-andy/20news-test-vectors/part-r-00000", classOf[Writable],
classOf[VectorWritable], minPartitions = 10).map(t => (t._1, t._2.get()))
mahout> val keyVec = rdd.map(_._1).collect.distinct
keyVec: Array[org.apache.hadoop.io.Writable] =
Array(/comp.os.ms-windows.misc/9141, /comp.sys.mac.hardware/52007,
/rec.autos/101620, /rec.sport.baseball/104334, /sci.crypt/15200,
/sci.electronics/54486, /sci.space/61469, /talk.politics.guns/54503,
/talk.politics.mideast/77353, /talk.religion.misc/84570)
mahout> keyVec.size
res1: Int = 10
{code}
however I'm expecting several more disticnt values for each category eg.:
/comp.os.ms-windows.misc/9141
/comp.os.ms-windows.misc/9142
{...}
{code}
mahout seqdumper -i /tmp/mahout-work-andy/20news-test-vectors/part-r-00000 |
less
{code}
shows the first entry is for: Key: /alt.atheism/51119 which doesn't seem be
showing up at all in the keys read in from the SparkContext.
was (Author: andrew_palumbo):
actually Looks like the like above is not where everything is getting dropped
looking closer in the mahout spark-shell I several other Keys (though not all):
{code}
mahout> val rdd = sdc.sequenceFile(path =
"/tmp/mahout-work-andy/20news-test-vectors/part-r-00000", classOf[Writable],
classOf[VectorWritable], minPartitions = 10).map(t => (t._1, t._2.get()))
mahout> val keyVec = rdd.map(_._1).collect.distinct
keyVec: Array[org.apache.hadoop.io.Writable] =
Array(/comp.os.ms-windows.misc/9141, /comp.sys.mac.hardware/52007,
/rec.autos/101620, /rec.sport.baseball/104334, /sci.crypt/15200,
/sci.electronics/54486, /sci.space/61469, /talk.politics.guns/54503,
/talk.politics.mideast/77353, /talk.religion.misc/84570)
mahout> keyVec.size
res1: Int = 10
{code}
however I'm expecting several more disticnt values for each category eg.:
/comp.os.ms-windows.misc/9141
/comp.os.ms-windows.misc/9142
{...}
{code}
mahout seqdumper -i /tmp/mahout-work-andy/20news-test-vectors/part-r-00000 |
less
{code}
shows the first entry is for: Key: /alt.atheism/51119 which doesn't seem be
showing up at all in the keys read in from the SparkContext.
> 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)
