Data Access 2003 PJDB.htm

J

John Sitka

The data access document for Microsoft Project 2003 mentions the
Transact/SQL scripts for Project database creation and stored procedures.

Where are these located? What are the file extensions?


Thnaks
 
R

Rob Schneider

I recall seeing them on the Project CD (sorry, don't have the CD here to
check that).

Hope this is useful to you. Let us know.

rms
 
J

John Sitka

Thanks Rob, unfortunately I don't have the DVD here
to check so that must be it.

I've been reading some of your past posts. Good stuff.
It's been a real battle to concieve of a workable production
floor schedule over the last couple months. When the folks "in
the know" here really sat down and thought it out they realized
they need to distribute the update of "actuals."

I see that this has been done before,

Can you make a recommendation on having maybe 100
projects inserted in a master project resident in an SQL database.

will it work? I test drove Project Server, first impressions
were it has too many layers of "Company noise" on top of it that
I don't think systems needs to carry around with us.
But if the data design is so much improved over Project 2003
I might consider it. We may eventually move to a slight resource
manager model but that is a long way away and there is so much
value for us in getting distributed actuals that I don't want to complicate
things.

For example: I will probably run the actuals collected from
the enterprise to a separate table then schedule an SQL job
to batch update the Project actuals.
 
R

Rob Schneider

I guess my initial thought is to ask if it's really necessary to use the
SQL Server db just to store project data? Does that add unnecessary
complexity (and cost)? What's the benefit? What risks are you trying
to control? My hunch it's not worth it to use SQL db server, but I
don't have the benefit of knowing as much as you do about this.

I think Project Server is a terrific workgroup/enterprise project
management system. It's a 'big' project to get it going, though,
because it cuts into the core of how projects are managed by people
(remember that people manage projects, Project doesn't). For large
number of people in the same "enterprise" doing projects that have
resource or other organisational overlaps ... Project Server is
terrific. Just not simple. I feel it worth it, though, for larger
organisations.

Frankly, as a starting point, having 100 separate project mpp files
included into one (or more) master Project mpp files shouldn't be a
problem. Sounds ok to me. Try it. Simulate it by creating 100 (or
more) separate files and link them all together and see if it works.

What is a "distributed actual"?

Hope this is useful to you. Let us know.

rms
 
J

John Sitka

Great response Rob

SQL because it is plays into my skill set. I've read some stuff about well
VBA this OLE that,
or Access. No thanks. Not when SQL is an option.

"Distributed actuals" represent the need to include what we call our Value
Stream Managers to
be presented with a task list and then update the actual work for those
tasks. (I hope they can
signal tasks as complete as well.)

Truely the best way to think about this is how it actually goes down on the
shop floor.

The famous question " How much do you have left "

How many hours of work does the value stream manager have left to complete
that task.

Ideally we would like to ask him that every day and let him update his
tasks.
As he is a manager as well, he may have some flexability in terms of how he
orders his tasks,
whether they are part of the critical path or not or parallel tasks. That is
often dependent on the
custom nature of a build and it's varities.

This concept is in conflict with what many view as scheduling. But in truth
the root cause of any ability
to move forward in terms of scheduling and what we are trying to do in
project is that the Projects do not
reflect reality close enough, it lacks concurrency. By bringing the
production plan closer to real time we
can get the accuracy we need to use it as a finite scheduling system. Right
now by the time folks get around
to updating them they are out of date. I don't want to roll out a bar code
real time collection system to get this
done, it would be a total bloat.
(I just need a knife not a food processer, because the knife I can use in
the woods)

For example. Reschedule, level a few times, get a nice plan. then by lunch
something has happened
which demands one of our huge expensive finite resources is going to be
reassigned to a different
task, every production plan would has the opportunity to be impacted by
this. The task that got
bumped would show zero progress on a day one would expect to see 10 hours
actual. Our Project Manager,
manages the variance and rescehdules. His managing of the variance and
rescheduling is only as valuable
as the accuracy and concurrency of the actuals.
In order to recover the actuals from a Value Stream Manager everyday I
envision a task bar docked or web application
that the Value Stream managers fills in his best estimate on how much is
left. These guys know that daily.
The company thinks they know that a week in advance, sometimes they do
sometimes they don't.
The rate of change is too rapid and our buffers cannot be that big.

Thus he takes on the role of a status manager for his range of tasks he is
reponsible for.

Our shop manager can then have an upto date project to work with.

So really that is why SQL because I don't believe there is an Data Provider
for Project,
is there? That would be real funny. I swear I looked for one.
 
R

Rob Schneider

Humm... gotta think about all this. More than my brain can deal with at
this time of nigth while sipping a glass of wine ...

What immediately comes to mind ... if you haven't read the file
projdb.htm that is part of the install of Project, read it again (and
often). Describes the database structure and the hoops you must jump
through to do direct updates of the database. Remember that while you
can get the data into a SQL db, it doens't have the triggers and stored
procedures that you may expect that will "protect" the db integrity.
That protection/validation is all built into Project. Hence, you may
want to be looking at the application you want to write through the lens
of Project as the "front" end. Sorry, that will mean using Project VBA
.... but it's very powerful and simplifies a lot for you.

(I still have my doubts that you should be focusing on SQL db ... that
just seems a diversion to the real issue ... just a hunch).

Hope this is useful to you. Let us know.

rms
 
J

John Sitka

Wow, I guess I made a big assumption or a misread of PJDB.htm.

I'll for sure read it again but.

This part is where the A_S_S-U-M_E came in
It is recommended that the stored procedures included with project be used
to update this column with the required values

<list of tables and columns>

I guess I assumed those procs were my ticket.
So yeah, thanks for the heads up about the RI and validation.
So a thousand miles of hacked RI in SQL or VBA.
I've been at this fork in the road before!

You've been a great help Rob, I know my description below
dosen't mix with wine. It's a pretty disjointed set.
 
R

Rob Schneider

Well ... maybe you have something I haven't read yet. I can't find this
text in 2002 projdb.htm. Using 2003 version? I wasn't aware of any
special SP's being availble other than scripts for creation and some
SP's for backing up data. I'm now curious. I'm a bit biased. I'd
still look to licking this via Project rather than reinvention.
 
J

John Sitka

Just little pieces of the puzzle then
How do I distribute task/dispatch sheets?
On the web or application.
 
J

John Sitka

Yes 2003 version....
but like you say stay calm, read, read and reread the PJDB.htm.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top