I just noticed that

SET SESSION;

is not only correct SQL syntax, it is also described as correct syntax
(e.g. following error E_US1019).

What should I expect it to do? It is hard to test for its effects (if
any).

--
Roy

UK Ingres User Association Conference 2011 will be on Tuesday June 7 2011.
Put the date in your diary today.

Your right Roy. I've tried this on everything I've got back to a VMS II2.0 box and they all say its cool.

Looks like a hole in the syntax parser. I'd suggest you bring that one to the next code sprint.

Marty

-----Original Message-----
From: Roy Hann [mailto:specially (AT) processed (DOT) almost.meat]
Sent: 24 September 2010 08:41
To: info-ingres (AT) kettleriverconsulting (DOT) com
Subject: [Info-Ingres] What does SET SESSION; do?

I just noticed that

SET SESSION;

is not only correct SQL syntax, it is also described as correct syntax
(e.g. following error E_US1019).

What should I expect it to do? It is hard to test for its effects (if
any).

--
Roy

UK Ingres User Association Conference 2011 will be on Tuesday June 7 2011.
Put the date in your diary today.


_______________________________________________
Info-Ingres mailing list
Info-Ingres (AT) kettleriverconsulting (DOT) com
http://ext-cando.kettleriverconsulti...fo/info-ingres

Martin Bowes wrote:

It *looks* like a hole, but it is actually also *described* as valid:

E_US1019 line 1, Syntax error on 'EOF'. The correct syntax is:
SET SESSION
[ADD|DROP PRIVILEGES ( priv {,priv} )]
[,READ ONLY|WRITE]
[,ISOLATION LEVEL isolation_level]
[WITH [SECURITY_LABEL = 'label' | INITIAL_LABEL]
[,ON_ERROR = ROLLBACK TRANSACTION|STATEMENT]
[,ON_USER_ERROR = NOROLLBACK | ROLLBACK TRANSACTION]
[,ON_LOGFULL = ABORT|COMMIT|NOTIFY]
[,[NO]PRIVILEGES [= ( priv {,priv} )| ALL | DEFAULT]]
[,[NO]PRIORITY [= nnnn | INITIAL | MAXIMUM | MINIMUM]]
[,[NO]DESCRIPTION [= description]]
]
(Fri Sep 24 11:01:32 2010)

OK, maybe the parser and the error message are both working from the
same excessively permissive grammar. Or maybe that syntax really does
do something.

--
Roy

UK Ingres User Association Conference 2011 will be on Tuesday June 7 2011.
Put the date in your diary today.

On Sep 24, 11:05*am, Roy Hann <specia... (AT) processed (DOT) almost.meat> wrote:

Resets everything back to the defaults after you've changed them with
earlier "set session" statements?
(just a guess...)

On Sep 24, 2010, at 3:41 AM, Roy Hann wrote:

It doesn't do anything. I think the grammar is a side effect of there
being basically two kinds of SET SESSION:
SET SESSION thing, thing
SET SESSION WITH thing, thing. There's two sets of rules
and either or both can be empty. There's no good reason for the
two forms that I know of, it just happened.

Karl

Tim Ellis wrote:

That's the sort of plausible possibility I was expecting to hear, yes.

--
Roy

UK Ingres User Association Conference 2011 will be on Tuesday June 7 2011.
Put the date in your diary today.

On Sep 24, 2010, at 6:37 AM, Roy Hann wrote:

Plausible, but it happens to not be the case. I don't think there is any
way to set session default (although it's maybe not a bad idea!).
I'd do it with set session default, though, not just set session.

Karl

On Sep 24, 2010, at 6:38 AM, Karl Schendel wrote:

Friday's on comp.databases.ingres are always so riveting. I'm just glad
nobody brought up the release of Postgres 9 ...

;-)

Mikey

Mike Leo wrote:

Now you've brought it up you can clean it up too. I don't want to step
in it.

--
Roy

UK Ingres User Association Conference 2011 will be on Tuesday June 7 2011.
Put the date in your diary today.

Well, one could consider it as a kind of a valid "no-op" operation,
which would probably reset some session's runtime information
(iierrornumber, iirowcount, etc.)


--
Bodo