[
https://issues.apache.org/jira/browse/OAK-9653?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Martin Schulze updated OAK-9653:
--------------------------------
Description:
I've experienced this behavior und the following circumstances:
* Query with an index tag option ("option (index tag x)")
* The index the tag has an regex property with REGEX_ALL_PROPS
What happens is that the resulting query to lucene, which can be inspected in
[1] is:
{noformat}
+:ancestors:/test +:indexTag:x{noformat}
So the {{:indexTag}} will be added to the Lucene index as a property, because
{{:indexTag}} matches the regex, which can be inspected in [2]
I've added a failing unit test for this in [0], where I just copied the
\{{descendantTest}} and added the IndexTag
I've also tried to come around with a custom regex which excludes every node
which starts with a colon ( regex: {{{}"^[^:\\/][^\\/]*$"{}}}). This leads to
another problem where the index won't be selected and the query will traverse.
[0] [https://github.com/apache/jackrabbit-oak/pull/448/files]
[1]
[https://github.com/apache/jackrabbit-oak/blob/trunk/oak-lucene/src/main/java/org/apache/jackrabbit/oak/plugins/index/lucene/LucenePropertyIndex.java#L340]
[2]
[https://github.com/apache/jackrabbit-oak/blob/trunk/oak-search/src/main/java/org/apache/jackrabbit/oak/plugins/index/search/IndexDefinition.java#L1500]
was:
I've experienced this behavior und the following circumstances:
* Query with an index tag option ("option (index tag x)")
* The index the tag has an regex property with REGEX_ALL_PROPS
What happens is that the resulting query to lucene, which can be inspected in
[1] is:
{noformat}
+:ancestors:/test +:indexTag:x{noformat}
So the \{{:indexTag}} will be added to the Lucene index as a property, because
{{:indexTag}} matches the regex, which can be inspected in [2]
I've added a failing unit test for this in [0].
I've also tried to come around with a custom regex which excludes every node
which starts with a colon ( regex: {{"^[^:\\/][^\\/]*$"}}). This leads to
another problem where the index won't be selected and the query will traverse.
[0] [https://github.com/apache/jackrabbit-oak/pull/448/files]
[1]
[https://github.com/apache/jackrabbit-oak/blob/trunk/oak-lucene/src/main/java/org/apache/jackrabbit/oak/plugins/index/lucene/LucenePropertyIndex.java#L340]
[2]
https://github.com/apache/jackrabbit-oak/blob/trunk/oak-search/src/main/java/org/apache/jackrabbit/oak/plugins/index/search/IndexDefinition.java#L1500
> Adding the index tag option interferes with regex properties, leads to return
> zero results
> ------------------------------------------------------------------------------------------
>
> Key: OAK-9653
> URL: https://issues.apache.org/jira/browse/OAK-9653
> Project: Jackrabbit Oak
> Issue Type: Bug
> Components: indexing
> Affects Versions: 1.40.0
> Reporter: Martin Schulze
> Priority: Major
>
> I've experienced this behavior und the following circumstances:
> * Query with an index tag option ("option (index tag x)")
> * The index the tag has an regex property with REGEX_ALL_PROPS
> What happens is that the resulting query to lucene, which can be inspected in
> [1] is:
> {noformat}
> +:ancestors:/test +:indexTag:x{noformat}
> So the {{:indexTag}} will be added to the Lucene index as a property, because
> {{:indexTag}} matches the regex, which can be inspected in [2]
> I've added a failing unit test for this in [0], where I just copied the
> \{{descendantTest}} and added the IndexTag
> I've also tried to come around with a custom regex which excludes every node
> which starts with a colon ( regex: {{{}"^[^:\\/][^\\/]*$"{}}}). This leads to
> another problem where the index won't be selected and the query will traverse.
> [0] [https://github.com/apache/jackrabbit-oak/pull/448/files]
> [1]
> [https://github.com/apache/jackrabbit-oak/blob/trunk/oak-lucene/src/main/java/org/apache/jackrabbit/oak/plugins/index/lucene/LucenePropertyIndex.java#L340]
> [2]
> [https://github.com/apache/jackrabbit-oak/blob/trunk/oak-search/src/main/java/org/apache/jackrabbit/oak/plugins/index/search/IndexDefinition.java#L1500]
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
