D-Net
/
dnet-hadoop
13
4
Fork 10
Code Issues 14 Pull Requests 8 Projects 2 Releases 1 Wiki Activity

[cleaning] align parameter names

This commit is contained in:
Claudio Atzori 2022-12-20 21:43:59 +01:00
parent 5816ded93f
commit 33bdad104e
1 changed files with 6 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/CleanContextSparkJob.java
View File

@ -47,8 +47,8 @@ public class CleanContextSparkJob implements Serializable {
String inputPath = parser.get("inputPath");
log.info("inputPath: {}", inputPath);
String workingPath = parser.get("workingPath");
log.info("workingPath: {}", workingPath);
String workingDir = parser.get("workingDir");
log.info("workingDir: {}", workingDir);
String contextId = parser.get("contextId");
log.info("contextId: {}", contextId);
@ -67,12 +67,12 @@ public class CleanContextSparkJob implements Serializable {
isSparkSessionManaged,
spark -> {
cleanContext(spark, contextId, verifyParam, inputPath, entityClazz, workingPath);
cleanContext(spark, contextId, verifyParam, inputPath, entityClazz, workingDir);
});
}
private static <T extends Result> void cleanContext(SparkSession spark, String contextId, String verifyParam,
String inputPath, Class<T> entityClazz, String workingPath) {
String inputPath, Class<T> entityClazz, String workingDir) {
Dataset<T> res = spark
.read()
.textFile(inputPath)
@ -106,11 +106,11 @@ public class CleanContextSparkJob implements Serializable {
.write()
.mode(SaveMode.Overwrite)
.option("compression", "gzip")
.json(workingPath);
.json(workingDir);
spark
.read()
.textFile(workingPath)
.textFile(workingDir)
.map(
(MapFunction<String, T>) value -> OBJECT_MAPPER.readValue(value, entityClazz),
Encoders.bean(entityClazz))