[ https://issues.apache.org/jira/browse/CALCITE-1254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15308769#comment-15308769 ]
Josh Elser commented on CALCITE-1254: ------------------------------------- Thanks again, Julian! bq. Yes, I suppose that was a little contentious. I figured that if someone was asking for at most N rows then clearly there could be no more than N rows in the first frame. Given that cap, I can't imagine a scenario where an existing client would see a change in behavior. My vote would be to clarify the documentation of prepareAndExecute to state that maxRowsInFirstFrame is capped by maxRowCount. Great. I can do that. bq. And by the way, don't use (int) to cast to int, use saturated cast, so that 2^32 becomes Integer.MAX_VALUE rather than 0, and importantly -1 remains -1. Right you are again. I'll make a pass over the codebase. I think there might have been some other instances of this. > Support PreparedStatement.executeLargeBatch > ------------------------------------------- > > Key: CALCITE-1254 > URL: https://issues.apache.org/jira/browse/CALCITE-1254 > Project: Calcite > Issue Type: Bug > Components: avatica > Reporter: Julian Hyde > Assignee: Josh Elser > Priority: Blocker > Fix For: avatica-1.8.0 > > > In CALCITE-1128 we added support for PreparedStatement.executeBatch. This > added ExecuteBatchResult with a field {{int[] updateCounts}}. > I think that field should have been {{long[]}} instead. Elsewhere we have > been converting update counts from {{int}} to {{long}}, in line with changes > to the JDBC API. > If changing this field from {{int[]}} to {{long[]}} will be a breaking change > we should consider halting 1.8 to get this in. -- This message was sent by Atlassian JIRA (v6.3.4#6332)