Re: DBA_LOCK fields

oh goody, maybe you can answer a question for me. I've been to oracle support, oracle developer program, contacts at oracle, etc. and had no joy.

say I update a row in table fred. there's now a row lock on that row, and an entry in v$lock that points to fred. say I now update a second row in fred. there's still only one entry in v$lock. my question is, since the kernel can keep track of individual row locks, where can I find out which rowid is locked by which session? I tried to make sense of x$kglck, or whatever it's called, but it seemed to also have only one row.

if I can get an answer to this, I can get rid of a process that runs under dbms_job and keeps a register of locks as advised by applications over database pipes. don't tell me to use dbms_lock, as that is full of commits, which doesn't suit

I'm running 7.3.3.3

ta in advance

-- 
Andrew - Wizzard

barnetaj_at_bp.com

Kristiaan J. Kolk <akolk_at_ix.netcom.com> wrote in article
<351155EE.834AFC89_at_ix.netcom.com>...


> No,

> 

> The lock_id1 and lock_id2 have different logical meanings depending on

> the lock type. for example if lock type is TM, id1 is the object id. If

> type is TX, id1 means the rbs nr & slot nr, id2 is the wrap number.

> 

> So as you can see it all depends !

> 

> Anjo.

> 

> I work for Oracle, but I speak for myself.

> 

> Dalibor Kovac wrote:

> 

> > Can anybody tell me what do the fields

> >   LOCK_ID1

> >   LOCK_ID2

> > in the DBA_LOCK view contain?

> > I know that one of those IDs is an object_id of a locked object,

> > but what's the other one?

> >

> > Thanks,

> >

> > Dalibor Kovac

> > dalibor_kovac_at_yahoo.com

> 

> 

> 

>