[HUDI-68] Pom cleanup & demo automation (#846)
- [HUDI-172] Cleanup Maven POM/Classpath - Fix ordering of dependencies in poms, to enable better resolution - Idea is to place more specific ones at the top - And place dependencies which use them below them - [HUDI-68] : Automate demo steps on docker setup - Move hive queries from hive cli to beeline - Standardize on taking query input from text command files - Deltastreamer ingest, also does hive sync in a single step - Spark Incremental Query materialized as a derived Hive table using datasource - Fix flakiness in HDFS spin up and output comparison - Code cleanup around streamlining and loc reduction - Also fixed pom to not shade some hive classs in spark, to enable hive sync
This commit is contained in:
@@ -217,12 +217,23 @@ public class CompactionCommand implements CommandMarker {
|
||||
final String sparkMemory,
|
||||
@CliOption(key = "retry", unspecifiedDefaultValue = "1", help = "Number of retries")
|
||||
final String retry,
|
||||
@CliOption(key = "compactionInstant", mandatory = true, help = "Base path for the target hoodie dataset")
|
||||
final String compactionInstantTime) throws Exception {
|
||||
@CliOption(key = "compactionInstant", mandatory = false, help = "Base path for the target hoodie dataset")
|
||||
String compactionInstantTime) throws Exception {
|
||||
boolean initialized = HoodieCLI.initConf();
|
||||
HoodieCLI.initFS(initialized);
|
||||
|
||||
if (HoodieCLI.tableMetadata.getTableType() == HoodieTableType.MERGE_ON_READ) {
|
||||
if (null == compactionInstantTime) {
|
||||
// pick outstanding one with lowest timestamp
|
||||
Option<String> firstPendingInstant = HoodieCLI.tableMetadata.reloadActiveTimeline()
|
||||
.filterCompletedAndCompactionInstants().filter(instant -> instant.getAction()
|
||||
.equals(HoodieTimeline.COMPACTION_ACTION)).firstInstant().map(HoodieInstant::getTimestamp);
|
||||
if (!firstPendingInstant.isPresent()) {
|
||||
return "NO PENDING COMPACTION TO RUN";
|
||||
}
|
||||
compactionInstantTime = firstPendingInstant.get();
|
||||
}
|
||||
|
||||
String sparkPropertiesPath = Utils.getDefaultPropertiesFile(
|
||||
scala.collection.JavaConversions.propertiesAsScalaMap(System.getProperties()));
|
||||
SparkLauncher sparkLauncher = SparkUtil.initLauncher(sparkPropertiesPath);
|
||||
|
||||
Reference in New Issue
Block a user