The current project uses the data set development model and three-tier structure.
In fact, regardless of whether it is three-tiered or not, the biggest benefit of using data source control is: in the past, multi-table related queries were required, but now many of them can be eliminated. The data source control will automatically help you do it;) Isn’t it very relaxing and enjoyable?
For example, table A has three foreign keys, ID1, ID2, ID3. You only need to convert the three fields into templates, and then select the appropriate controls to bind to the three data source controls. OK, you don’t have to worry about the rest.
Secondly, objdatasource among the data source controls is indeed easy to use. Other data source controls are either too simple or specialized (sitemap). Only with the cooperation of the three-layer structure can the power of the data source control be fully exerted. What I can write by hand, I haven't encountered anything that objDs can't do yet. But you need to "change your brain" and switch your thinking. The implementation is indeed not the same as writing code by hand. I now have a page of code. In the past, other people's project background used more than 1K lines, but now I only use less than 400 lines. This cannot but be said that objDs still has advantages in some aspects.
In addition, when building an adapter, you should pay attention to whether the generated data type is consistent with the database. In particular, the char(1) type is generally set to byte. If the code is correct but an error occurs, this is often the problem. Just modify it and it will be OK
http://www.cnblogs.com/emilchan/archive/2006/11/30/578033.html