alessia.bardi
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

allow to speficy the target hive DB name in the infospace import workflow

This commit is contained in:
Claudio Atzori 2019-11-07 17:38:09 +01:00
parent a52d5bde4f
commit 5308f05a02
4 changed files with 20 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/graph/SparkGraphImporterJob.java
View File

@ -21,25 +21,30 @@ public class SparkGraphImporterJob {
.builder()
.appName(SparkGraphImporterJob.class.getSimpleName())
.master(parser.get("master"))
.config("hive.metastore.uris", parser.get("hive.metastore.uris"))
.config("hive.metastore.uris", parser.get("hive_metastore_uris"))
.enableHiveSupport()
.getOrCreate();
final JavaSparkContext sc = new JavaSparkContext(spark.sparkContext());
final String inputPath = parser.get("sourcePath");
final String hiveDbName = parser.get("hive_db_name");
spark.sql(String.format("CREATE DATABASE IF NOT EXISTS %s", hiveDbName));
// Read the input file and convert it into RDD of serializable object
GraphMappingUtils.types.forEach((name, clazz) -> {
final JavaRDD<Tuple2<String, String>> inputRDD = sc.sequenceFile(inputPath + "/" + name, Text.class, Text.class)
.map(item -> new Tuple2<>(item._1.toString(), item._2.toString()));
spark.createDataset(inputRDD
.filter(s -> s._1().equals(clazz.getName()))
.map(Tuple2::_2)
.map(s -> new ObjectMapper().readValue(s, clazz))
.rdd(), Encoders.bean(clazz))
.limit(1000)
.write()
.mode(SaveMode.Overwrite)
.saveAsTable("openaire." + name);
.saveAsTable(hiveDbName + "." + name);
});
}

3
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/graph/input_graph_parameters.json
View File

@ -1,5 +1,6 @@
[
{"paramName":"mt", "paramLongName":"master", "paramDescription": "should be local or yarn", "paramRequired": true},
{"paramName":"s", "paramLongName":"sourcePath", "paramDescription": "the path of the sequencial file to read", "paramRequired": true},
{"paramName":"h", "paramLongName":"hive.metastore.uris","paramDescription": "the hive metastore uris", "paramRequired": true}
{"paramName":"h", "paramLongName":"hive_metastore_uris","paramDescription": "the hive metastore uris", "paramRequired": true},
{"paramName":"db", "paramLongName":"hive_db_name", "paramDescription": "the target hive database name", "paramRequired": true}
]

6
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/graph/oozie_app/config-default.xml
View File

@ -16,7 +16,11 @@
<value>spark2</value>
</property>
<property>
<name>hive.metastore.uris</name>
<name>hive_metastore_uris</name>
<value>thrift://iis-cdh5-test-m3.ocean.icm.edu.pl:9083</value>
</property>
<property>
<name>hive_db_name</name>
<value>openaire</value>
</property>
</configuration>

7
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/graph/oozie_app/workflow.xml
View File

@ -4,6 +4,10 @@
<name>sourcePath</name>
<description>the source path</description>
</property>
<property>
<name>hive_db_name</name>
<description>the target hive database name</description>
</property>
<property>
<name>sparkDriverMemory</name>
<description>memory for driver process</description>
@ -36,7 +40,8 @@
<spark-opts>--executor-memory ${sparkExecutorMemory} --executor-cores ${sparkExecutorCores} --driver-memory=${sparkDriverMemory} --conf spark.extraListeners="com.cloudera.spark.lineage.NavigatorAppListener" --conf spark.sql.queryExecutionListeners="com.cloudera.spark.lineage.NavigatorQueryListener" --conf spark.sql.warehouse.dir="/user/hive/warehouse"</spark-opts>
<arg>-mt</arg> <arg>yarn-cluster</arg>
<arg>--sourcePath</arg><arg>${sourcePath}</arg>
<arg>--hive.metastore.uris</arg><arg>${hive.metastore.uris}</arg>
<arg>--hive_db_name</arg><arg>${hive_db_name}</arg>
<arg>--hive_metastore_uris</arg><arg>${hive_metastore_uris}</arg>
</spark>
<ok to="End"/>
<error to="Kill"/>