[
https://issues.apache.org/jira/browse/CALCITE-523?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14971679#comment-14971679
]
Josh Elser commented on CALCITE-523:
------------------------------------
We currently break up results into multiple batches via {{Frame#done}}, but we
have no size-based limiting.
Could repurpose this issue or close as done and open a new one specific to
size-based result batching.
> Incremental fetch in remote JDBC driver
> ---------------------------------------
>
> Key: CALCITE-523
> URL: https://issues.apache.org/jira/browse/CALCITE-523
> Project: Calcite
> Issue Type: Bug
> Components: avatica
> Reporter: Julian Hyde
> Assignee: Julian Hyde
>
> The remote JDBC driver currently returns all rows in its initial response. It
> should break the result up into batches.
> Batch algorithm could be as complex as "fetch 100 rows, or 100K bytes, which
> is larger".
> Requests PrepareAndExecuteRequest, ExecuteRequest (not yet created) and
> metadata requests CatalogsRequest etc. all return a ResultSetResponse.
> ResultSetResponse contains a "rows" field. Need to add fields { int
> firstRowOrdinal; int rowCount; boolean mayHaveMoreRows; }, and a new
> FetchRequest { int rowOrdinal; minRowCount; int minByteCount; }
> FetchRequest returns a ResultSetResponse whose signature is null.
> If mayHaveMoreRows is false, we are definitely done. If it is true, there may
> or may not be more rows. A subsequent fetch might be empty. An empty fetch or
> mayHaveMoreRows = false indicates end of data.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
