Database Discussion Boards

I would like to thank everyone in advance for their answers as I am extremely stuck. Also apologies if I put this question on wrong as it is my first attempt at this.

I am currently developing a C# application for a project at work, which uses a SQL 2008 server as a back end database. I know it’s an older way but I have been using windows forms and ADO.Net and have managed through this site and various books to get quit far but now have hit a problem that I cannot find answered on here. Some tables are going to be very large and hence I want to implement paging and have done this through a stored procedure. The application works out what page it requires and passes the start row index and number of rows to return to the procedure. This works perfectly (though I have read various posts to do with performance, and different ways of achieving paging in SQL).
My problem has arisen now that I won’t to introduce searching into the mix. Most posts/articles are geared towards ASP.net application which a user specifies the search criteria and the results for that search are returned in a page form. But I want to copy the functionality of a system we currently run at work called Syspro. Syspro will return you to the page relevant to the data you have just edited, added, or searched for. For example it opens the form with a data grid view on page one, I add new data which will be 100 pages in, after adding the data and closing the add form it finds the page that result is on and shoots the data grid view page to that page and high lights the new row. This works in the same way when you search and edit as you can skip through data on the add/mod form as well.

Is there a way of replicating this behavior in SQL or is this something that is handled in the C# application. I need a way of determining what page the data I require is on, but I understand that SQL only assigns row numbers after a query is run against that table. I have thought of using the primary key to save me the trouble, but some tables will allow deletes.

Lance Parker wrote:
Is there a way of replicating this behavior in SQL o

Your question about what "behavior" you want to replicate is confusing but yes you can implement paging in SQL.

Basic steps.
1. Create your SQL with appropriate search parameters. You should still use parametrized values.
...a. If you have just a couple of items you can hard code the SQL.
...b. With more than a couple items create the SQL dynamically.
2. Apply appropriate ordering to the SQL. This can include creating it dynamically.
3. Apply paging (dynamic SQL again) See the following link

SQL Server 2005 Paging Results[^]

At the business level you might also want to consider what happens to paging if the data being paged changes underneath. Very common answers are
1. It won't, so there is no problem.
2. It won't in relation to the user because the work flow and usage insures that.

Be VERY sure that you analyze this at the business level before deciding that this could be a problem.

Also you should require sufficient search constraints to insure that no single result has lot of pages. Three pages is ok, 100 isn't. If someone claims otherwise then note that users do not randomly seek data. They know what they are looking for. So make them tell you it. Paging is a convenience so they don't have to tell you everything in detail but ignoring it can inhibit rather than help workflow.

SQL

Thank you jschell.
    But I must apologise as after returning to work today and looking at the programme in which I am trying to copy some functionality from, I have realised I haven’t explained what I am after properly, and fear I have caused some confusion. But as for what you have told me in the above it validates most of the code I have written, thank you for that.

    With that I’ll try again to explain myself, I caused confusion by saying there was a search involved "Apologies". In a nut shell what I am after is basically "Selecting and displaying a page based on a records location". Now I know that until a query is run against the table there is no theoretical position of a record. But what the programme does is open the pageForm on page one of the data in that table. Although there is a ID column on the table (as I have poked around in the programme files) the page is organised by the first column which is stock-Code. Hence you can page through all the results from a stock code beginning with "A" right through to "Z" which I think is around probably 1000 pages at the minute, not that we go through them all. Then a user opens a new form to edit or add a new record, when the form is closed and the data returned to the database the pageform will display the page relevant to the stock code added, and you can page backwards and forwards from this page. The same applies to the search, which is conducted in a separate searchForm, which displays the possible results for that criteria, which the user selects one, and returns to the pageForm which now displays the page with that record on.

Do you jschell or any one have any idea of how this is accomplished?

It is ordered by the stock code.

Everything I said at the SQL level applies.

You would write code in a db layer to do that.
Then the GUI level would use the db layer.