Hi,

We are looking to map about 200+ entities. These equate to about 10 logical (domain) models with entities that have some overlap with some of the other logical (domain) models.

1) How does LLBL Gen handle such large context - performance? Can these be broken out into multiple logical models in the designer with one edmx - if performance is not likely to be hit. Or I guess we would have to attach entities from one context to another?

2) Do you have any examples of STE use with LLBLGen?

3) How can one test STE implementations in "unit" tests - SQLCE, SQLLite? does the generated code from LLBLGen allow one to cater for this - or do we have to make the necessary adjustments to the generated output to handle such testing - any examples?

4) Can we apply both TPT and TPH strategies within a single model? I guess if this were to be possible, it would need to know what descriminator field it should use for TPH.

Rik wrote:

1) How does LLBL Gen handle such large context - performance? Can these be broken out into multiple logical models in the designer with one edmx - if performance is not likely to be hit. Or I guess we would have to attach entities from one context to another?

LLBLGen Designer is capable to manage hundreds of entities. Whether the EntityFramework can handle this is another story. Anyway, in you LLBLGen project properties there is one that is used to determine how you want to group your entities:

**GroupUsage **This setting controls how the grouping functionality of the designer is used with respect to the generated output: solely as visual grouping construct (default) (so all elements in the entire project are seen as one project) or as separate projects (one per group). If you choose to see groups as separate projects, validation will verify that there are no ties between groups whatsoever and will report errors if there are. If you choose to see groups as a visual grouping construct only, validation will verify if there are duplicate named elements among groups and will report errors if there are.

Rik wrote:

2) Do you have any examples of STE use with LLBLGen? ... 3) How can one test STE implementations in "unit" tests - SQLCE, SQLLite? does the generated code from LLBLGen allow one to cater for this - or do we have to make the necessary adjustments to the generated output to handle such testing - any examples?

LLBLGen generates ready-to-use EF code, what you do with is up to you. So any EF example out there should work. If you want to see some videos of how to work with EF and LLBLGen you can see them here: http://llblgen.com/Pages/videos.aspx

Rik wrote:

4) Can we apply both TPT and TPH strategies within a single model? I guess if this were to be possible, it would need to know what descriminator field it should use for TPH.

Yes you can. Here are some relevant documentation links about that: - Concepts - Entity Inheritance - How to add/edit inheritance information