[
https://issues.jboss.org/browse/TEIID-5072?page=com.atlassian.jira.plugin...
]
Steven Hawkins commented on TEIID-5072:
---------------------------------------
The proposal for this would look like:
{code}
create view v1 (i integer, k integer not null auto_increment primary key) OPTIONS
(UPDATABLE true)
as select x, y from tbl;
create trigger on v1 instead of insert as
for each row begin atomic
insert into tbl (x) values (new.i); key.k = cast(generated_key('y') as
integer);
end;
{code}
This introduces a new variable group key, which would only be in scope for the insert
trigger on a view with an auto_increment primary key - and only for the key columns.
There would be a new function introduced called generated_key(string) that could fetch the
first generated key column value from the last insert statement or null if there were no
matching columns, etc.
Generated keys should be passed from triggers
---------------------------------------------
Key: TEIID-5072
URL:
https://issues.jboss.org/browse/TEIID-5072
Project: Teiid
Issue Type: Sub-task
Components: Query Engine
Reporter: Steven Hawkins
Assignee: Steven Hawkins
Fix For: 10.0
In the context of an update trigger we should be able to convey the generated keys to the
caller. In the case where only a single insert occurs, this should be automatic.
Otherwise we'll need functions/variables to get/set the generated keys in the
procedure. For example see the LAST_INSERT_ID function on mysql.
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)