Accepting enhancement suggestions for new releases

Thank you. Your suggestions and feedback are always welcome! If you have new ideas for further improvements, please don't hesitate to port them in this support forum.

We have many stored procs with hundreds and sometimes thousands of lines each. I'm wondering if there is a way of storing bookmarks inside the stored proc. Maybe if the bookmarks that are created in the usual way are then written with special codes as comments near the start of the procedure. (This could also double as a 'table of contents' for the stored proc). Then when the procedure is later opened SQL Assistant recognises the codes in the special comments and 'rebuilds' the bookmarks automatically for that proc.

This would provide a permanent navigation system for procedures, where broad sections of a procedure could be automatically bookmarked and shown in the 'table of contents' at the top of the procedure for quick navigation.

As an aside, it would be useful to be able to assign a name/label to each bookmark rather than just a number.

This is a very interesting idea. I have shared it with the development team and waiting for their feedback.

Thank you for your input.

One frustration I have with SQL Server is debugging code that contains variables (which typically is most code!). Often you can't execute a block of code insitu (by highlighting, then F5) because the variables are declared earlier in the code and also they require values to be assigned. So you need to then copy the code out of the proc to another query, manually add declarations, and then replace @variables with hardcoded values. Then if you made changes to the code as part of debugging you need to copy it back into the original procedure, remembering to replace the hardcoded values with the original variable names.

Is there a way that SQL Assistant will let you highlight a block of code and when you execute it, automatically check for declarations for any variables earlier in the proc (before the highlighted code) and then display a dialog box listing all the variables in the highlighted code, the expected data types based on the earlier declarations, and a textbox next to each to allow a value ot be enterred. I know Oracle tools such as TOAD do this. It would save a lot of time when debugging, as you can execute code insitu regardless of whether it contains variables or not.

In the current version it doesn't support such feature. To the best of my knowledge you are the first one asking for it?

However, if you copy and paste the block of code into a new edit window, it will tell you which variables are missing. Also, if in the procedure code you place mouse pointer over a variable name it will display a hint with the variable declaration and if you click this hint it will scroll the editor and highlight the actual declaration so you can quickly copy or modify it. You can also use one click, always visible bookmarks feature if you need to jump back to the place where you were before (or use Ctrl+Z - the smart undo feature to move to the last edit point but not chnage it until you see the place). While all of that is not exactly what you want, I think it does a great deal of easing the pain.

Thanks for that. Those tips are defintely handy, and I'll be using them. But you'd be amazed how useful it is to be able to just highlight some code, execute it, be prompted for the values for any variables (which you enter) and then have it run, in place. It really does save a lot of time if you do a lot of testing of code fragments.

SQL Server gives you a rowcount in the bottom right hand corner of the results screen, but that's about it. It would be useful to be able to click on a field heading in a result set and if its a numerical column, get a Sum of all the values (and maybe Min, Max, Avg etc) same as Excel gives you. Also perhaps a flag indicating whether at least one NULL is present. Could restrict if necessary to only rowcounts < a certain feasible size. This would be useful to quickly check for expected totals on the fly and would avoid copy/paste to excel to check this type of stuff, or having to include extra SQL code to report it. There seems to be a lot of space avaialble for this info in the status bar where SQL Server already shows the rowcount.

Thanks. This is an interesting idea to have quick totals for selected numeric columns, but in SQL Assistant's result grid, I mean Ctrl+Shift+F9 results.

PS. SQL Assistant is not specific to any editor and cannot modify behavior of controls outside of the editor box.

May I suggest Softtree to make some required enhancements to tooltip such that it would display tips in Visul Studio editior style.
For example:
1) Instead of showing sp/fun parameter list vertically, show it horizontally.
2) When I use S.proc or function it should highlight current parameter in bold.

Hitesh

Thank you for your suggestions. This is a very significant enhancement indeed which we are planning in future versions. Please note that as compared to strict programming languages (C#, C++, Java, etc…) function parameters in SQL can specified both by position or by name and in the latest case (which ism ore common) they can be specified in any order, so that strictly there is no such thing as a current parameter. On top of that Oracle supports overloaded functions and procedures (same name, different parameters). This is really a non-trivial task.

Being a developer I understand your point of view, but initially you can start with system defined function and stored procedure. As a practice, developer does not change Para info when coding although such flexibility is available in most databases. Give it a try at least for system defined sp & fun!

Hitesh

Thanks again. We already have this enhancement in the queue for implementing Visual Studio stile function and procedure tips, but the implementation schedule for this enhancement is not yet finalized.

As you are aware that Visual Studio 2010 (Beta 1) is out and it has all new code editor build using WPF. Will SQL Assit get integrated with new WPF editor window class?

Hitesh

Thanks. I personally wasn't aware of new v10. I'll and alert to dev team so they know about it.

By the way, previous betas and CTP versions turned very "unstable" in the sense that Microsoft has made numerous significant changes in the design before the public release. WE will likely wait for the final CTP before working on the integration stuff.

Regards.

This is pretty minor I guess but the placement of the red underline indicating where a problem in the code is can be misleading and a bit confusing. It quite often highlights the valid line of code AFTER the actual error, rather than any part of the actual error line itself. I can see why it happens in some cases (e.g. where you have missed some code and it tries to interpret your next bit of code as part of the previous statement), but if this next bit of code is on a new line (and often seperated by a blank row!) then chances are the problem lies before it. With duplicate declare statements for example the error appears at the start of the valid declare statement that FOLLOWS the duplicate. This is a bit confusing.

If the underlined error was always placed at the SQL word just BEFORE the bad one, or on the bad word itself only if it is not on a new line then this might be more intuitive for debugging.

Just a thought really.
Glenn