[
https://issues.apache.org/jira/browse/ASTERIXDB-3289?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ian Maxon updated ASTERIXDB-3289:
---------------------------------
Labels: triaged (was: )
> MurmurHash128Bit can read the wrong field data
> ----------------------------------------------
>
> Key: ASTERIXDB-3289
> URL: https://issues.apache.org/jira/browse/ASTERIXDB-3289
> Project: Apache AsterixDB
> Issue Type: Bug
> Components: OTH - Other
> Affects Versions: 0.9.9
> Reporter: Wail Y. Alkowaileet
> Assignee: Wail Y. Alkowaileet
> Priority: Major
> Labels: triaged
> Fix For: 0.9.9
>
>
> The hash computer (namely
> [MurmurHash128Bit|https://github.com/apache/asterixdb/blob/0a25d5620e56c728954e5abc813c3ba513e4859a/hyracks-fullstack/hyracks/hyracks-storage-am-bloomfilter/src/main/java/org/apache/hyracks/storage/am/bloomfilter/impls/MurmurHash128Bit.java#L132C58-L132C58])
> reads dummy field data as follows:
> {noformat}
> tuple.getFieldData(DUMMY_FIELD) {noformat}
> PKs in columnar tuples are stored in different buffers. Thus, the current
> implementation of
> [MurmurHash128Bit|https://github.com/apache/asterixdb/blob/0a25d5620e56c728954e5abc813c3ba513e4859a/hyracks-fullstack/hyracks/hyracks-storage-am-bloomfilter/src/main/java/org/apache/hyracks/storage/am/bloomfilter/impls/MurmurHash128Bit.java#L132C58-L132C58]
> will always read the data of the first PK even if the second PK is required.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
