added the outputPath parameter and the reset path to remove the outputath directory

This commit is contained in:
Miriam Baglioni 2020-05-11 16:10:24 +02:00
parent 5ab3424c77
commit e883daf87e
1 changed files with 7 additions and 2 deletions

View File

@ -16,6 +16,10 @@
<name>sourcePath</name> <name>sourcePath</name>
<description>the source path</description> <description>the source path</description>
</property> </property>
<property>
<name>outputPath</name>
<description>the path were to store the graph without the blacklisted relations</description>
</property>
</parameters> </parameters>
<start to="reset-outputpath"/> <start to="reset-outputpath"/>
@ -25,7 +29,8 @@
<action name="reset-outputpath"> <action name="reset-outputpath">
<fs> <fs>
<delete path='${workingDir}/blacklist'/> <delete path="${outputPath}"/>
<mkdir path="${outputPath}"/>
</fs> </fs>
<ok to="read_blacklist"/> <ok to="read_blacklist"/>
<error to="Kill"/> <error to="Kill"/>
@ -87,7 +92,7 @@
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
</spark-opts> </spark-opts>
<arg>--sourcePath</arg><arg>${sourcePath}/relation</arg> <arg>--sourcePath</arg><arg>${sourcePath}/relation</arg>
<arg>--outputPath</arg><arg>${workingDir}/relation</arg> <arg>--outputPath</arg><arg>${outputPath}/relation</arg>
<arg>--hdfsPath</arg><arg>${workingDir}/blacklist</arg> <arg>--hdfsPath</arg><arg>${workingDir}/blacklist</arg>
<arg>--mergesPath</arg><arg>${workingDir}/mergesRelation</arg> <arg>--mergesPath</arg><arg>${workingDir}/mergesRelation</arg>
</spark> </spark>