cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ajay (JIRA)" <j...@apache.org>
Subject [jira] [Created] (CASSANDRA-9770) Expose API to load prepare statement based on preparedId
Date Thu, 09 Jul 2015 10:45:11 GMT
Ajay created CASSANDRA-9770:
-------------------------------

             Summary: Expose API to load prepare statement based on preparedId
                 Key: CASSANDRA-9770
                 URL: https://issues.apache.org/jira/browse/CASSANDRA-9770
             Project: Cassandra
          Issue Type: Bug
          Components: Drivers (now out of tree)
         Environment: Cassandra 2.0.16
Java driver 2.1.6
Ubuntu 15.04
            Reporter: Ajay
             Fix For: 2.1.6


Currently whenever client calls session.prepare(Statement), the Java driver prepares the statement
with cassandra. Cassandra prepares the statement, caches in memory and returns prepareId.
The driver caches it in the session and uses it subsequently. 

In case, we have another session (as in web application or REST services, there are multiple
JVM and so different cluster and session instances), currently we need reprepare the same
query for which we need the original query string and other details (like consistency level
and so on).

Since cassandra is already caching in memory (also a ticket is there to persist as well in
3.x), we can have an API to load the preparestatement from any session (if not there in the
cache already) as below

Session.loadPreparedStatement(MD5Digest id) 

where the driver can call Cassandra and load the preparedstatement (to get metadata, resultSetMetadata,
routingKeyIndexes etc...).

WIth this, the client need to cache only the prepareId and can avoid multiple round trip to
prepare the query (same query) again with different hosts.




--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message