I have some similar issue sometime ago.
I am running MAHOUT_08_SNAPSHOT.
I found out my problem is in the data set I downloaded.
After unzipping dataset, there are some files starting with ., ending with
gz are not really in gz format
eg: commons.apache.org/user/._200312.gz,.....
If you have the same situation, then
either
dropping all files starting with . using script.
or
modify file selection function at SequenceFilesFromMailArchives.java
line 97
public boolean accept(File current) { --line 97
-------add block begins
if (current.getName().startsWith(".")){
//System.out.println("skip " +current.getName() );
return false;
}
------add block ends
if (current.isDirectory()) {
......}
}
Then everything is fine.
Hope this helps.
Sam
On Wed, Jun 27, 2012 at 7:12 PM, Josh Patterson <[email protected]> wrote:
> I've seen this same thing happen. It should work on the directory of
> .gz files in the asf-mail archives.
>
> On Fri, Jun 15, 2012 at 5:33 PM, Lance Norskog <[email protected]> wrote:
> > Do these use the same directory structure as the apache mail example?
> >
> > On Thu, Jun 14, 2012 at 11:09 PM, Tian, Phil <[email protected]>
> wrote:
> >> Hi,
> >>
> >> When I try to create sequence files from email archives with the below
> command, I always get nothing. Can anyone give me some hint on this issue?
> Appreciate your help.
> >>
> >> [training@rassi5prcimf-vm trunk]$ bin/mahout
> org.apache.mahout.text.SequenceFilesFromMailArchives --charset "UTF-8"
> --subject --body --input /tmp/mahout-work-training/asf-email/content
> --output
> /tmp/mahout-work-training/asf-output/classification/bayesian/seq-files
> >> MAHOUT_LOCAL is not set; adding HADOOP_CONF_DIR to classpath.
> >> Running on hadoop, using HADOOP_HOME=/usr/lib/hadoop
> >> No HADOOP_CONF_DIR set, using /usr/lib/hadoop/conf
> >> MAHOUT-JOB:
> /home/training/workspace/trunk/examples/target/mahout-examples-0.7-SNAPSHOT-job.jar
> >> Picked up _JAVA_OPTIONS: -Xmx1024m
> >> Picked up _JAVA_OPTIONS: -Xmx1024m
> >> 12/06/15 14:01:00 WARN driver.MahoutDriver: No
> org.apache.mahout.text.SequenceFilesFromMailArchives.props found on
> classpath, will use command-line arguments only
> >> 12/06/15 14:01:02 INFO text.SequenceFilesFromMailArchives: Parsed 0
> messages from /tmp/mahout-work-training/asf-email/content in time: 7
> >> 12/06/15 14:01:02 INFO text.SequenceFilesFromMailArchives: Conversion
> took 1676ms
> >> 12/06/15 14:01:02 INFO driver.MahoutDriver: Program took 1700 ms
> (Minutes: 0.028333333333333332)
> >>
> >> Best regards,
> >>
> >> Phil Tian
> >> Enterprise Architect
> >> Intel IT Labs
> >>
> >>
> >>
> >
> >
> >
> > --
> > Lance Norskog
> > [email protected]
>
>
>
> --
> Twitter: @jpatanooga
> Principal Solution Architect @ Cloudera
> hadoop: http://www.cloudera.com
>
