Turns out the documentation is correct. In the past I have had trouble with this and thus I posted my solution. Although it remains true that you can just make the subfolder its own application this might lead to the subfolder not inheriting the parents web.config settings. This is something you might want. Good information!

It is my understanding that the web.config applies to each application. If you have a subfolder in which you want its own config then you will need to make it its own application within IIS. This usually has little impact on the application but will possibly have some deployment issues since the root of the application will be in a different memory space than the sub folder application. So your answer is no (because they will need to become seperate applications and you used that word specifically). If you can make the subfolder its own application using IIS or VS.NET then yes ;) http://www.jasncab.com/huberblog :: Jason Huber

You should have used DTS or the Upsizing Wizard. Both are well documented and easy to use. http://www.jasncab.com/huberblog :: Jason Huber

I use 10G regularly in two environments installed on Windows 2003. Works just fine. ASP.NET and the driver you currently use will work fine. http://www.jasncab.com/huberblog :: Jason Huber

Sure I have a video out there that I made for my CIS 410 class. Let me know if you have problems seeing it. http://www.jasncab.com/devry/cis410 The idea is that you fill the dataset then bind to the Dim ds As DataSet Dim strSQL As String Dim strResult As String strSQL = "select rollid, rollid || '-' || description || ' ' || rolldate as descdate from qa_rollouts " 'FILL THE DATASET REMOVED _ ADD BACK IN cboRollouts.DataTextField = "descdate" cboRollouts.DataValueField = "rollid" cboRollouts.DataSource = ds cboRollouts.DataBind() ds.Dispose() ds = Nothing Really with anything in form load you want to make sure that you wrap your binding code inside of a ispostback. If you dont you will either overwrite what was there or add to it. Usually it isnt a good idea to leave the ispostback off. Also enabling viewstate has caused confusing depending on what you expect. http://www.jasncab.com/huberblog :: Jason Huber

Yes you can really close to what you have already. You are joining from the sub out. and that should be ok. Its called a correlated sub query.. The problem is I dont see your point in this particular example. It could be that I am just missing it. But really you could just rewrite your example SELECT tb1.namedfields, AS tb2_ID, tb2.Author AS tb2_Author, tb2.Address AS tb3_Address FROM tb1 LEFT JOIN tb2.author ON FROM tb2.id = tb1.id left join tb3 on tb2.Address = tb3.ID http://www.jasncab.com/huberblog :: Jason Huber

I think you are taking normalization too far. Dont concern yourself as to storing addresses more than once. The address in each table I create is usually just an attribute. This means I would have an address for each company and person. Same with phone. Although it is possible that a company and person might have the same address or phone it isnt important that they are the same physical record in some table. "Owning" a company - you work for it. I see three tables in your entire post. companies -> companies_people <- people This way people can be part of a company or NOT. Company table would have a main phone, fax etc and address(s) People would have their own ID, address and phone. YES you might have duplicated data but it certainly could not be called a transitive dependancy. http://www.jasncab.com/huberblog :: Jason Huber

I dont think it is simple... but you could certainly try something like: Here is how you would do it using NW: select a.employeeid, count(b.employeeid) from employees a, employees b where a.employeeid = b.reportsto group by a.employeeid What we have is a self referencing table. What we need to do is essentially add the table twice in the from clause with two seperate aliases and then correlate the two in the where. I should be using ANSI standard joins here, but it was quick. so for you we need something like select a.id, count(b.id) from T1 a, T1 b where a.id = b.parentid group by a.id; Let me know if this works. http://www.jasncab.com/huberblog :: Jason Huber

I run into this all the time in my CIS349 and CIS 405 classes. In fact I just emailed a student of mine about a similar problem. use the oracle to_date or to_char functions. in SQL: select * from orders where orderdate between to_date('01/23/2005', 'MM/DD/YYYY') AND to_date('02/23/2005', 'MM/DD/YYYY') for the reverse.... select to_char(orderdate, 'MM/DD/YYYY') from orders; In the latter version when you get your data back from the db into your data container you will already have it in the right format. http://www.jasncab.com/huberblog :: Jason Huber