[MDEV-15457] Pipe function technique Created: 2018-03-03  Updated: 2018-08-13  Resolved: 2018-08-13

Status: Closed
Project: MariaDB Server
Component/s: Server
Fix Version/s: N/A

Type: Task Priority: Major
Reporter: Tadas BalaiĊĦis Assignee: Unassigned
Resolution: Duplicate Votes: 0
Labels: None

Issue Links:
PartOf
is part of MDEV-8100 Table functions (aka SQL functions re... Stalled

 Description   

I'd like to get an advice, how can I achieve pipe function (like Oracle) functionality in MariaDB. I know I can create procedure that returns recordset, that's fine. But the only way is to fill some temporary table and issue select at the end of procedure. That means that full content of recordset should be prepared and then directed to the output. What I'm looking for is possibility to return recordset data gradually: take for example first 100 rows, make additional calculation/aggregation inside the procedure and put the output to the caller. And next 100 rows would be processed only if calling side would issue subsequent fetches. So procedure could be ready to output millions of rows only if caller continue fetching, but it also would be ready to output first rows fast enough. Is there such a possibility?



 Comments   
Comment by Sergei Golubchik [ 2018-08-13 ]

This kind of result streaming was considered for MDEV-8100, but it's not implemented yet.

Generated at Thu Feb 08 08:21:27 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.