On 2/4/21 7:29 PM, Mark Rotteveel wrote:
On 2021-02-04 15:55, Norbert Saint Georges wrote:
Mark Rotteveel a écrit :
On 2021-02-04 09:39, Norbert Saint Georges wrote:
yes, this is what I found but there is no definition RDB $ FIELD_TYPE
= 30 or 31 in rdb $ types in FB4, what is it for?
The EXTENDED TIMESTAMP WITH TIME ZONE and EXTENDED TIME WITH TIME
ZONE are protocol-only data-types, where it will always include the
time zone offset, so client applications without named time zone
support can determine the offset. See also SET BIND in the Firebird
4 RC-1 release notes.
So if I understand correctly, I will never risk finding the values
32748 or 32750 in OO api (firebird.pas)?
No, your understanding is not correct. You will be able to see those
values in the OO API
To be precise - one can meet them in XSQLDA too.
if you have a connection created with isc_dpb_set_bind, or where the
SET BIND statement was used, to define a mapping to the EXTENDED time
zone types (eg isc_dpb_set_bind=TIME ZONE TO EXTENDED, or SET BIND OF
TIME ZONE TO EXTENDED)
Given these are protocol-only data types
"protocol-only" is not what exactly takes place - in embedded
connection extended timezone will also be present. That datatypes (and
that's how SET BIND works in general) arrive when one requests
IMessageMetadata from engine - i.e. asks engine to describe statement.
, they can occur in the description of data types of columns and
parameters of statements only. They cannot occur in the metadata
tables (the RDB$* tables).
Yes. They are never used inside engine.
But please add ticket re adding them to rdb$types - I think it's better
to have them here, at least for reference
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel