www.otn.oracle.com ? This goes no where.
"Jim Kennedy" <kennedy-down_with_spammers_at_attbi.com> wrote in message
news:T0Aoa.12552$Si4.6065_at_rwcrnsc51.ops.asp.att.net...
> Fix the application.
> Oracle!=SQLServer. Read about Oracle's locking and multi versioning
scheme
> in the docs. (www.otn.oracle.com, sign up for free and access all the docs
> on line). At worst I could see a select for update statement. I think
the
> problem is you are trying to get around SQLServer's limitations.
> Don't do the same thing in Oracle you do in SQLServer; doesn't make sense
> usually.
> Jim
>
> --
> Replace part of the email address: kennedy-down_with_spammers_at_attbi.com
> with family. Remove the negative part, keep the minus sign. You can
figure
> it out.
> "Bob Kilmer" <rprgrmr_at_yahoo.com> wrote in message
> news:h7zoa.56244$D15.1616373_at_twister.tampabay.rr.com...
> > My question concerns locks.
> >
> > I am a developer new to a VB6/Oracle 9i client-server application and
> > somewhat new to Oracle although I have developed against SQL Server and
> > Access quite a bit. One to several (3,4,...,10?) clients located on
> separate
> > machines use this Oracle database - no more. No other application or
> process
> > out of our control accesses this DB.
> >
> > The practice has been to explicitly lock particular rows in particular
> > tables under certain circumstances - while some input is pending, for
> > instance. Other clients trying to perform similar operations on the same
> > data may try to lock the same rows, only to be turned away. This
'turning
> > away" is acceptable - it just means that other clients should not
> interfere
> > with what is going on with those data and the data locking those records
> > implies.
> >
> > The problem arises in managing the code and the locks. In the course of
> some
> > user input process at the client GUI, many sub routines might need to
> access
> > the data that will have been locked when the input process was
initiated.
> > The locks are released and reinstated repeatedly as each bit of code
does
> > its thing. The code is not well modularized, so often code drops a lock,
> > anticipating a need for data access by subsequent code somehwere else,
or
> > reinstates a lock, anticipating that it should in case the other
spaghetti
> > code forgot to, or whatever - who knows? There are occasionally module
> level
> > or global booleans, but IMHO, this just adds to the quagmire.
> >
> > Can the client set a lock on Oracle db objects that is only a lock for
> > others?
> >
> > Wouldn't it be just as effective if each client set a flag in some table
> or
> > other that, in effect said, "do not disturb - I am working here", then
> > cleared it when done?
> >
> > Something like this would help make this "lock-unlock-lock-unlock-(hope
no
> > one jumps in!)-lock?-lock?-unlock?-..." nightmare go away.
> >
> > Help, please!
> >
> > Bob Kilmer
> >
> >
> >
>
>
>
Received on Sun Apr 20 2003 - 12:55:13 CDT
