[
https://issues.apache.org/jira/browse/RYA-266?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16140331#comment-16140331
]
ASF GitHub Bot commented on RYA-266:
------------------------------------
Github user DLotts commented on a diff in the pull request:
https://github.com/apache/incubator-rya/pull/149#discussion_r135073929
--- Diff:
extras/indexing/src/main/java/org/apache/rya/indexing/accumulo/temporal/AccumuloTemporalIndexer.java
---
@@ -158,16 +167,20 @@ public Configuration getConf() {
private void storeStatement(final Statement statement) throws
IOException, IllegalArgumentException {
// if the predicate list is empty, accept all predicates.
// Otherwise, make sure the predicate is on the "valid" list
- final boolean isValidPredicate = validPredicates.isEmpty() ||
validPredicates.contains(statement.getPredicate());
+ final boolean isValidPredicate = validPredicates == null ||
validPredicates.isEmpty() || validPredicates.contains(statement.getPredicate());
if (!isValidPredicate || !(statement.getObject() instanceof
Literal)) {
return;
}
+
final DateTime[] indexDateTimes = new DateTime[2]; // 0 begin, 1
end of interval
extractDateTime(statement, indexDateTimes);
if (indexDateTimes[0]==null) {
return;
}
+ if (!this.isInit)
--- End diff --
Indexes should not create their own batchwriter because it does not get
flushed properly.
If we call init() here, then the batchwriter may not have been set. If it
did not initialize at this point, then this is a bug and should fail.
> Temporal index fails to write when using non-mock accumulo
> ----------------------------------------------------------
>
> Key: RYA-266
> URL: https://issues.apache.org/jira/browse/RYA-266
> Project: Rya
> Issue Type: Sub-task
> Components: dao, sail
> Environment: indexing
> Reporter: David W. Lotts
> Assignee: David W. Lotts
> Priority: Blocker
> Labels: flush, index, temporal
> Original Estimate: 24h
> Remaining Estimate: 24h
>
> Temporal indexing writes nothing to it's table name: [prefix]temporal.
> It does seem to work for a mock instance. It might also work for a large
> ingest since it would force a flush.
> See the parent RYA-72 for an explanation.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
