forked from D-Net/dnet-hadoop
changed the way to save the results on filesystem
This commit is contained in:
parent
87f802821e
commit
8f12292daa
|
@ -100,7 +100,7 @@ public class SparkCountryPropagationJob2 {
|
|||
broadcast_datasourcecountryassoc).as(Encoders.bean(ResultCountrySet.class));
|
||||
|
||||
if(writeUpdates){
|
||||
writeUpdates(potentialUpdates.toJavaRDD(), outputPath + "/update_" + resultType);
|
||||
writeUpdates(potentialUpdates, outputPath + "/update_" + resultType);
|
||||
}
|
||||
|
||||
if(saveGraph){
|
||||
|
@ -147,9 +147,13 @@ public class SparkCountryPropagationJob2 {
|
|||
log.info("Saving graph table to path: {}", outputPath);
|
||||
//log.info("number of saved recordsa: {}", new_table.count());
|
||||
new_table
|
||||
.toJavaRDD()
|
||||
.map(r -> OBJECT_MAPPER.writeValueAsString(r))
|
||||
.saveAsTextFile(outputPath , GzipCodec.class);
|
||||
.toJSON()
|
||||
.write()
|
||||
.option("compression", "gzip")
|
||||
.text(outputPath);
|
||||
// .toJavaRDD()
|
||||
// .map(r -> OBJECT_MAPPER.writeValueAsString(r))
|
||||
// .saveAsTextFile(outputPath , GzipCodec.class);
|
||||
|
||||
}
|
||||
|
||||
|
@ -219,9 +223,14 @@ public class SparkCountryPropagationJob2 {
|
|||
.map(value -> OBJECT_MAPPER.readValue(value, DatasourceCountry.class), Encoders.bean(DatasourceCountry.class));
|
||||
}
|
||||
|
||||
private static void writeUpdates(JavaRDD<ResultCountrySet> potentialUpdates, String outputPath){
|
||||
potentialUpdates.map(u -> OBJECT_MAPPER.writeValueAsString(u))
|
||||
.saveAsTextFile(outputPath, GzipCodec.class);
|
||||
private static void writeUpdates(Dataset<ResultCountrySet> potentialUpdates, String outputPath){
|
||||
potentialUpdates
|
||||
.toJSON()
|
||||
.write()
|
||||
.option("compression", "gzip")
|
||||
.text(outputPath);
|
||||
// map(u -> OBJECT_MAPPER.writeValueAsString(u))
|
||||
// .saveAsTextFile(outputPath, GzipCodec.class);
|
||||
}
|
||||
|
||||
|
||||
|
|
Loading…
Reference in New Issue