diff options
author | Ellen Spertus <ellen.spertus@gmail.com> | 2018-06-03 16:54:08 -0700 |
---|---|---|
committer | Ellen Spertus <ellen.spertus@gmail.com> | 2018-06-03 16:54:08 -0700 |
commit | a9fe176a305b474050727d80c1f42941ecf1a2dc (patch) | |
tree | 0f2596b134c169d69742b12d3c1defe6949b231b /scalding/src/main/scala/sandcrawler/HBaseRowCountJob.scala | |
parent | 5da6d85175992fae019e1d4d1ebeb4accfec1820 (diff) | |
download | sandcrawler-a9fe176a305b474050727d80c1f42941ecf1a2dc.tar.gz sandcrawler-a9fe176a305b474050727d80c1f42941ecf1a2dc.zip |
Added HBaseBuilder.build() and had HBaseRowCountJob call it.
Diffstat (limited to 'scalding/src/main/scala/sandcrawler/HBaseRowCountJob.scala')
-rw-r--r-- | scalding/src/main/scala/sandcrawler/HBaseRowCountJob.scala | 10 |
1 files changed, 3 insertions, 7 deletions
diff --git a/scalding/src/main/scala/sandcrawler/HBaseRowCountJob.scala b/scalding/src/main/scala/sandcrawler/HBaseRowCountJob.scala index 79ebbb1..9d074ab 100644 --- a/scalding/src/main/scala/sandcrawler/HBaseRowCountJob.scala +++ b/scalding/src/main/scala/sandcrawler/HBaseRowCountJob.scala @@ -24,13 +24,9 @@ class HBaseRowCountJob(args: Args) extends JobBase(args) with HBasePipeConversio } object HBaseRowCountJob { - def getHBaseSource = new HBaseSource( - //"table_name", - //"quorum_name:2181", + 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?) - new Fields("key"), - List("file"), - List(new Fields("size", "mimetype")), - sourceMode = SourceMode.SCAN_ALL) + List("file:size", "file:mime"), + SourceMode.SCAN_ALL) } |