v2.0 Beta2 lists excessive table entries

Aug 31, 2010 at 3:11 PM

Hello,

I just downloade and installed the v2.0 Beta 2 tool.  After I my DB Connection String and connected to the database (Oracle 10G), the "Select a table" picklist contained an excessive amount of tables.  In prior versions of the tool, the picklist only contained tables located in the user_tables table.  In this version, it looks like it contains all the tables from the all_tables table.  The difference is well over 1000 extra table references in my case.  This made it very time-consuming to find the tables I was looking for.

Was this the intended action for the new version?

Thanks - Peter

Developer
Aug 31, 2010 at 3:26 PM

Yes, it was intended. However, we no longer query from user_tables since it seems that most people (including myself) would get no tables. Instead, we now query from all_tables, sorry for your hassle. If I have a few minutes to spare tonight, I will add a new drop down with all owners so you can chose which tables to see. It's either that or modify the query to exclude the following owners: SYS, SYSTEM. Pretty much exclude anything that has to do with the system tables. Which is what I'm assuming your seeing in the thousands.

Also, it's not really recommended to generate your models on a per table basis. The reasoning is that NMG will not do a dependency check. So if you select to model TableA only and make reference to TableB and TableC. The mapping will still generate your foreign key properties and one-to-many's.

Aside from the lack of Oracle sequences (which is rather trivial to add manually for the time being). How did you find the new tool?

Sorry for the inconvenience.

Aug 31, 2010 at 4:32 PM

Yeah, it would be helpful if maybe there were another picklist that just had the user_tables table values.  For me, that is where my tables are.  On this topic, you mentioned something about not generating my models on a per table basis.  I am new to O/R mapping but I was somewhat surprised that when it came to generate the files, I could only select a single table.  This made me think that the tool would not be able to see the relationships.  When I talked to a friend who uses Hibernate with Java, his O/R mapping file generator allows for him to select multiple tables.  I have him my two tables in this test schema and he used his tool to produce the two Hibernate mappings files and Java class files.  All the referecnes seemed to be there.  So my confusion is understanding how else I would generate my models.  Based on your statement, I need to do something different and this is what I don't understand.  How do you recommend that I generate the models?

As for the new tool version, it does do the one-to-many relationship stuff so that is great.  The interface itself is easy to use and pretty intuative.  I have found two things:

1) I get a compile warning on the generated .cs files because there are two "using System;" references.

2) The sequences defined for my tables (for the primary key) are not listed when I select my table from the picklist.  This causes a problem in the mapping file definition for the primary key.

- Peter