SQL Power Software Forum

Hi,
Since we're on the subject of wannas...

I would like to be able to exclude a table from the Forward Engineer and DM Compare functions.

Specifically my issue (today!) is that I have a couple of legacy tables (in the same database but a different schema) to which my new schema must link via foreign keys. As Architect doesn't yet support cross-database linking, I have been forced to create a mock table in the new schema and link to it so that my foreign keys are preserved. Of course, I don't actually want the mock table in the database, so I have been manually editing out the CREATE statement that Forward Engineer generates.

It shouldn't be too hard to add a "Skip During FE or DM Compare" attribute to a table or ...why not...to columns as well?

Keep up the good work!

Here is what I would like to see:

1) sub-models so that I can split a large model into smaller pieces but still be able to have relationships between the tables

2) support for check constraints

3) a customizable auto-layout. The current auto-layout is nice (for reverse engineering) but leaves too much spaces between the tables (for my personal taste)

4) To be able to move tables even with a high zoom. I have a model with > 300 tables and when using "Zoom to fit" I can't move the tables anymore because they are so small I can't grab Title bars

5) support for data domains (I think this is already on the roadmap)

6) Separation of the layout and the actual data model in the XML file to make it easier to identify changes when mutliple users are working on the same file (e.g. through subversion). I have already posted this, but I though it might be a good idea to list it here again.

Thanks for the work done so far!

It shouldn't be too hard to add a "Skip During FE or DM Compare" attribute to a table or ...why not...to columns as well?

This is a good idea. Pretty simple solution to your problem, and flexible enough to solve other problems we haven't though of yet.

Vote for bug 1593 to get this elevated.

-Jonathan

1) sub-models so that I can split a large model into smaller pieces but still be able to have relationships between the tables

Yes, we want to go here too. I think bug 1593 will be the interim solution.

2) support for check constraints

I guess it depends on your SQL style, but check constraints have the same problem as views: they're a place where you have to deal with platform specific syntax. There are several ways to solve this, but we haven't really got a framework in place yet. I want to support check constraints and views within the same framework. Any ideas you have on this would be most welcome.

3) a customizable auto-layout. The current auto-layout is nice (for reverse engineering) but leaves too much spaces between the tables (for my personal taste)

Fair enough. There are several knobs we could expose.

4) To be able to move tables even with a high zoom. I have a model with > 300 tables and when using "Zoom to fit" I can't move the tables anymore because they are so small I can't grab Title bars

This you can already do. When you hold down the alt key, your cursor changes to a "move" cursor, and the whole table area becomes draggable.

5) support for data domains (I think this is already on the roadmap)

Yes. This will be the next large feature we tackle.

6) Separation of the layout and the actual data model in the XML file to make it easier to identify changes when mutliple users are working on the same file (e.g. through subversion). I have already posted this, but I though it might be a good idea to list it here again.

We've used GUIDs that don't change over time in the Wabit project files, and this has made those files much more mergeable. I think we'll do this in Architect too. It should be a backward-compatible change.

As for putting the layout in a completely separate file, I remember we had a discussion about this a while back, but I can't remember the conclusion. I think my main concerns were around how to keep the two in sync, but you may have addressed these.

Thanks for your input. It's very valuable and always appreciated!

-Jonathan

Thanks for the tip for dragging the table, makes live a lot easier ;)

I have seen on the main product page that you are planning "Server-based offering" and "Data Model Collaboration" for the 1.0 release (scheduled for the end of the year?).

Would that mean that several people could work concurrently on the same model? (because e.g. the model is stored in the database and not in an XML file)

That would be a very interesting feature for us.
I guess that will then be a paid option ("Enterprise vs. Community Edition")?

Yes on all counts:

* We're planning an "enterprise server" product that will be sold for some nonzero price (pricing TBD)

* When connected to this enterprise server, the Architect will act as a client which allows several people to work on the same model simultaneously

* The current plan is to make the enterprise server product available by the end of this year.

-Jonathan

That is good news

Thanks for the feedback

Views. I would like to be able to model my views with the ability to enter the code that creates the view. Also should include the reverse engineering of views and forward engineering as well. Maybe differentiate the look of the view on the model too.

In Erwin we have a subject area feature and is this available in this tool ? if Yes how to access it . I really appreciate the quick response. thanks.

Need support for EMC's new Greenplum database. It was built from PostgreSQL database.