aboutsummaryrefslogtreecommitdiffstats
path: root/scald-mvp/src/main/scala
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@archive.org>2018-05-22 16:37:21 -0700
committerBryan Newbold <bnewbold@archive.org>2018-05-24 00:02:36 -0700
commit29e4a83ff76da07bc6ad5d3f49d746ee0bc72023 (patch)
tree87008116fb431849a19e0ab06d4757ebe12c1208 /scald-mvp/src/main/scala
parent58447ff883db0331bfd2450b8607ba332f53b3b6 (diff)
downloadsandcrawler-29e4a83ff76da07bc6ad5d3f49d746ee0bc72023.tar.gz
sandcrawler-29e4a83ff76da07bc6ad5d3f49d746ee0bc72023.zip
fix up HBaseRowCountTest
Again, seems like test fixture must match *exactly* or very obscure errors crop up.
Diffstat (limited to 'scald-mvp/src/main/scala')
-rw-r--r--scald-mvp/src/main/scala/sandcrawler/HBaseRowCountJob.scala4
1 files changed, 4 insertions, 0 deletions
diff --git a/scald-mvp/src/main/scala/sandcrawler/HBaseRowCountJob.scala b/scald-mvp/src/main/scala/sandcrawler/HBaseRowCountJob.scala
index 162f729..5df6b2e 100644
--- a/scald-mvp/src/main/scala/sandcrawler/HBaseRowCountJob.scala
+++ b/scald-mvp/src/main/scala/sandcrawler/HBaseRowCountJob.scala
@@ -17,9 +17,13 @@ class HBaseRowCountJob(args: Args) extends JobBase(args) with HBasePipeConversio
val output = args("output")
val hbs = new HBaseSource(
+ //"table_name",
+ //"quorum_name:2181",
"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.GET_LIST, keyList = List("sha1:K2DKSSVTXWPRMFDTWSTCQW3RVWRIOV3Q", "sha1:C3YNNEGH5WAG5ZAAXWAEBNXJWT6CZ3WU"))
.read
.debug