On 15/01/13 16:01, Martin J. Evans wrote:
On 15/01/13 15:04, Charles Jardine wrote:
Are you saying that, in the case of a NULL variable, the indicator
variable does not indicate nullity?
I am saying when you have:
procedure fred(pcur OUT SYS_REFCURSOR) as
begin
pcur := NULL;
end;
then the output parameter DBD::Oracle sees appears to be a valid Oracle
statement as you can call OCIAttrGet for OCI_ATTR_STMT_STATE and it
works. The following snippet illustrates this:
OCIAttrGet_stmhp_stat2(imp_sth, (OCIStmt*)phs->desc_h, &stmt_state, 0,
OCI_ATTR_STMT_STATE, status);
The desc_h in the parameters must be an OCIStmt * or surely this would
not work.
In this case the REF CURSOR variable in question is explicitly null.
I would expect the value returned via OCI to be accompanied by an
indicator variable with a value of -1,indicating a null value.
If this is the case, the value of the output variable is,
to quote the OCI manual, 'unchanged'. It should be ignored.
I am suggesting that the indicator variable should be tested before
the looking at the value. If indicator is -1, the value could be
the cursor returned by a previous execution of the same statement
handle.
--
Charles Jardine - Computing Service, University of Cambridge
c...@cam.ac.uk Tel: +44 1223 334506, Fax: +44 1223 334679
