[ 
https://issues.apache.org/jira/browse/CASSANDRA-21069?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Caleb Rackliffe updated CASSANDRA-21069:
----------------------------------------
    Description: {{TrieMemoryIndex}} presently does not support 
prefix/suffix/contains queries, even though prefix queries should be 
conceptually straightforward for a trie. (i.e. We only store postings for 
complete terms in the trie right now.) We should explore our options for this 
in implementation space. It should be possible to attach payloads anywhere in 
the trie and see those payloads during traversal. It's also possible to collect 
data while a sub-trie is being modified (per [~blambov]). The non-trivial 
decision we have to make is how many postings lists we store for internal nodes 
of the trie. (ex. Should we store postings for sub-trees every N levels and 
merge otherwise?)

> Prefix queries on Memtable-attached index
> -----------------------------------------
>
>                 Key: CASSANDRA-21069
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-21069
>             Project: Apache Cassandra
>          Issue Type: Improvement
>          Components: Feature/SAI
>            Reporter: Caleb Rackliffe
>            Priority: Normal
>
> {{TrieMemoryIndex}} presently does not support prefix/suffix/contains 
> queries, even though prefix queries should be conceptually straightforward 
> for a trie. (i.e. We only store postings for complete terms in the trie right 
> now.) We should explore our options for this in implementation space. It 
> should be possible to attach payloads anywhere in the trie and see those 
> payloads during traversal. It's also possible to collect data while a 
> sub-trie is being modified (per [~blambov]). The non-trivial decision we have 
> to make is how many postings lists we store for internal nodes of the trie. 
> (ex. Should we store postings for sub-trees every N levels and merge 
> otherwise?)



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to