[
https://issues.apache.org/jira/browse/NIFI-626?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14557381#comment-14557381
]
Joseph Witt commented on NIFI-626:
----------------------------------
Toivo
It is great that you're digging into this so much. I would say that we should
strive to support arbitrarily large result sets. The ideal scenario is to
stream the response from the DB to the flowfile outputstream. This requires
all parts of the design chain to support it; jdbc client, data formats, etc..
If this cannot be done then we want to be sure to document that well. Such
cases where large objects must be held in memory are the main sources of issues
with garbage collection and the like.
Thanks
Joe
> ExecuteSQL processor for executing arbitrary SQL queries
> --------------------------------------------------------
>
> Key: NIFI-626
> URL: https://issues.apache.org/jira/browse/NIFI-626
> Project: Apache NiFi
> Issue Type: Sub-task
> Reporter: Toivo Adams
> Assignee: Toivo Adams
>
> For example query can be:
> SELECT * FROM orders WHERE orderId = '${order.id}'
> where ${order.id} is FlowFile attribute.
> Result will serialized using Avro.
> Avro gives as possibility to have query Resultset metadata (column name and
> types) included in FlowFile.
> Also Avro should allow streaming.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
