[
https://issues.apache.org/jira/browse/IGNITE-4370?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16047895#comment-16047895
]
Igor Sapego commented on IGNITE-4370:
-------------------------------------
[~vozerov],
1) In the beginning there also was metadata in response to start request, which
I've removed afterwards. However, don't you think that SQL query caching may be
useful if there is large SQL request (few hungered bytes) and many parameters,
~1KK? In this case there is going to be significant overhead, don't you think
so?
2) That was for the consistency reasons, as if you can't make request with one
set of parameters why would you be able to do exactly the same request but with
2 parameter sets? However, on the second though, I think this is unnecessary
restriction. Will fix.
> ODBC: Implement DML operations with parameters in batch.
> --------------------------------------------------------
>
> Key: IGNITE-4370
> URL: https://issues.apache.org/jira/browse/IGNITE-4370
> Project: Ignite
> Issue Type: Task
> Components: odbc
> Affects Versions: 1.7
> Reporter: Igor Sapego
> Assignee: Igor Sapego
> Labels: important
> Fix For: 2.1
>
>
> Currently, if user wants for example to insert 20k records they need to call
> {{SQLExecute}} for the 20k times. More than that, internally, we transmit and
> execute the same SQL query 20k times. This is a huge overhead. We should only
> transfer sql query once, transfer all the parameters in a batch and then
> execute it once using some fast, possibly internal API.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
