diff options
| author | Ellen Spertus <ellen.spertus@gmail.com> | 2018-06-09 10:38:04 -0700 |
|---|---|---|
| committer | Ellen Spertus <ellen.spertus@gmail.com> | 2018-06-09 10:38:04 -0700 |
| commit | 3e9b927ba668f13f818d20962a6fb6f5783407df (patch) | |
| tree | 312308cbf2578f97f3b72a03c11eb6e5e05212dc /scalding/src/main/scala | |
| parent | 78f6f9bb555390050ecbf758f1f4d36fafa2a099 (diff) | |
| download | sandcrawler-3e9b927ba668f13f818d20962a6fb6f5783407df.tar.gz sandcrawler-3e9b927ba668f13f818d20962a6fb6f5783407df.zip | |
Refactored HBaseStatusCount{Job,Test}.scala.
Diffstat (limited to 'scalding/src/main/scala')
| -rw-r--r-- | scalding/src/main/scala/sandcrawler/HBaseStatusCountJob.scala | 27 |
1 files changed, 2 insertions, 25 deletions
diff --git a/scalding/src/main/scala/sandcrawler/HBaseStatusCountJob.scala b/scalding/src/main/scala/sandcrawler/HBaseStatusCountJob.scala index 0675efc..27b3177 100644 --- a/scalding/src/main/scala/sandcrawler/HBaseStatusCountJob.scala +++ b/scalding/src/main/scala/sandcrawler/HBaseStatusCountJob.scala @@ -1,28 +1,5 @@ package sandcrawler -import cascading.property.AppProps -import cascading.tuple.Fields -import com.twitter.scalding._ -import java.util.Properties -import parallelai.spyglass.base.JobBase -import parallelai.spyglass.hbase.{HBaseSource, HBasePipeConversions} -import parallelai.spyglass.hbase.HBaseConstants.SourceMode +import com.twitter.scalding.Args -class HBaseStatusCountJob(args: Args) extends JobBase(args) with HBasePipeConversions { - val output = args("output") - - HBaseStatusCountJob.getHBaseSource - .read - .fromBytesWritable(List('status)) -// .debug - .groupBy('status){group => group.size('count)} - .write(Tsv(output)) -} - -object HBaseStatusCountJob { - def getHBaseSource = HBaseBuilder.build( - "wbgrp-journal-extract-0-qa", // HBase Table Name - "mtrcs-zk1.us.archive.org:2181", // HBase Zookeeper server (to get runtime config info; can be array?) - List("grobid0:status"), - SourceMode.SCAN_ALL) -} +class HBaseStatusCountJob(args: Args) extends HBaseCountJob(args, "grobid0:status") |