Running SlopeOneAverageDiffsJob returned an empty part-r-00000.gz

[Steve Kuo <ku...@gmail.com>]

The input file squidly.csv is a test file that worked with
SlopeOneRecommender.java.  I ran the job with the following command:

hadoop jar /data/mahout/mahout-distribution-0.4/mahout-examples-0.4-job.jar
org.apache.mahout.cf.taste.hadoop.slopeone.SlopeOneAverageDiffsJob -i
squidly -o squidlyPredict

Unfortunately, file squidlyPredict/part-r-00000.gz is an empty .gz file.

What is the expected input format for SlopeOneAverageDiffsJob?

Squidly.csv is quite simple
1,101,1.0
1,102,0.5
1,103,0.2
2,101,1.0
2,103,0.5
2,104,0.2
3,101,0.2
3,102,0.4
3,103,1.0
3,104,0.4
4,102,0.9
4,103,0.4
4,104,0.5
5,101,0.4

Re: Running SlopeOneAverageDiffsJob returned an empty part-r-00000.gz

[Steve Kuo <ku...@gmail.com>]

Thanks Sean.  You are correctly.  I was removing empty and comment lines
from the input and missed putting it back to the input.

It is working now.

Re: Running SlopeOneAverageDiffsJob returned an empty part-r-00000.gz

[Sean Owen <sr...@gmail.com>]

There was almost certainly an error while running the job. Look through your
logs for signs of such a problem?

Sean
On 10 Feb 2011 19:51, "Steve Kuo" <ku...@gmail.com> wrote:
> The input file squidly.csv is a test file that worked with
> SlopeOneRecommender.java. I ran the job with the following command:
>
> hadoop jar
/data/mahout/mahout-distribution-0.4/mahout-examples-0.4-job.jar
> org.apache.mahout.cf.taste.hadoop.slopeone.SlopeOneAverageDiffsJob -i
> squidly -o squidlyPredict
>
> Unfortunately, file squidlyPredict/part-r-00000.gz is an empty .gz file.
>
> What is the expected input format for SlopeOneAverageDiffsJob?
>
> Squidly.csv is quite simple
> 1,101,1.0
> 1,102,0.5
> 1,103,0.2
> 2,101,1.0
> 2,103,0.5
> 2,104,0.2
> 3,101,0.2
> 3,102,0.4
> 3,103,1.0
> 3,104,0.4
> 4,102,0.9
> 4,103,0.4
> 4,104,0.5
> 5,101,0.4