Question Develop ER Diagram for Progress 4GL
- Thread starter Saran
- Start date
tamhas
ProgressTalk.com Sponsor
There is a plug-in for Enterprise Architect that does it. Of course, that is much, much more capable than Erwin, but you may not be interested in the capability. You can also look at http://www.oehive.org/ABL2UML , but that is even more yet.
Note that the intrinsic problem of building a diagram from the OpenEdge schema is that the schema does not include table to table relationships like some schemas do. This means that you need to establish those relationships by some other process. The plugin on PSDN has a "relationship guesser" based on looking for matching field names, but that has some obvious limitations including:
1. It won't find Customer.ID linking to Order.CustomerID and the like.
2. If you have Order.ShipToID and Order.BillToID it can't find both (or either)
3. It won't work on databases with table prefixes on field names.
4. It will blow up if you have commonly named audit fields like DateLastChanged.
Unfortunately, there is no source for the relationship guesser so you can't fix those.
ABL2UML does show you the empirical uses of each table and thus would provide the needed information, but there is no automation.
Note that the intrinsic problem of building a diagram from the OpenEdge schema is that the schema does not include table to table relationships like some schemas do. This means that you need to establish those relationships by some other process. The plugin on PSDN has a "relationship guesser" based on looking for matching field names, but that has some obvious limitations including:
1. It won't find Customer.ID linking to Order.CustomerID and the like.
2. If you have Order.ShipToID and Order.BillToID it can't find both (or either)
3. It won't work on databases with table prefixes on field names.
4. It will blow up if you have commonly named audit fields like DateLastChanged.
Unfortunately, there is no source for the relationship guesser so you can't fix those.
ABL2UML does show you the empirical uses of each table and thus would provide the needed information, but there is no automation.
santhosh22s
Member
I go with tamhas.
You could also try pCASE for openedge and DB Designer. Just remembered that we used both long back.
You could also try pCASE for openedge and DB Designer. Just remembered that we used both long back.
Marian EDU
Member
you might also try zamolxis erd - http://zamolxis.ganimede.ro/erd/
Rob Fitzpatrick
ProgressTalk.com Sponsor
There are various products out there that may help you. I believe our developers have used ER/Studio from Embarcadero. I've seen it briefly but I can't attest to its capabilities. We used it to connect to Progress DBs via ODBC and suck out the schema to build a data model.
As Thomas indicated, your results with this approach will depend a lot on the quality of your schema. But even on a poorly-designed DB where key fields are named differently across tables, it can give you a starting point from which to work. You can use such a tool to manually "wire up" the table relations based on developers' knowledge of the implicit model in the code or supporting documentation if any.
As Thomas indicated, your results with this approach will depend a lot on the quality of your schema. But even on a poorly-designed DB where key fields are named differently across tables, it can give you a starting point from which to work. You can use such a tool to manually "wire up" the table relations based on developers' knowledge of the implicit model in the code or supporting documentation if any.