From e7a540431019e6b5c877e90e8e66aadd850c2741 Mon Sep 17 00:00:00 2001 From: Mark Hale Date: Thu, 26 Oct 2023 00:12:50 +0100 Subject: [PATCH] Improve bulk load performance by using a combiner. --- .../src/main/java/com/msd/gin/halyard/tools/HalyardBulkLoad.java | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/src/main/java/com/msd/gin/halyard/tools/HalyardBulkLoad.java b/tools/src/main/java/com/msd/gin/halyard/tools/HalyardBulkLoad.java index ef8bcc5b2..2ea79a0eb 100644 --- a/tools/src/main/java/com/msd/gin/halyard/tools/HalyardBulkLoad.java +++ b/tools/src/main/java/com/msd/gin/halyard/tools/HalyardBulkLoad.java @@ -721,6 +721,7 @@ protected int run(CommandLine cmd) throws Exception { tableDesc = hTable.getDescriptor(); RegionLocator regionLocator = conn.getRegionLocator(tableDesc.getTableName()); HFileOutputFormat2.configureIncrementalLoad(job, tableDesc, regionLocator); + job.setCombinerClass(job.getReducerClass()); } try (Keyspace keyspace = HalyardTableUtils.getKeyspace(getConf(), conn, tableDesc.getTableName(), null, null)) { try (KeyspaceConnection ksConn = keyspace.getConnection()) {