It's interesting to jump back and forth between both programming models as lately I've been working on a decent size SQL 2K database (4M+ rows). Personally, I'm a Query Analyzer guy and I do all my SQL work there. I'm used to having IntelliSense in VS, which is annoying not to have in QA, but for some reason I live with 8pt fonts for queries, but I couldn't stand it in VS. Go figure. In any case here are my dream SQL Tool programming features:
DTS support for Importing XML
- I'm uploading a monster (multi-gigabyte) XML file into SQL 2K and you should see the stupid pet tricks I have to do to load it. I *don't* want to manually define the schema and create a mapping file, and a VBS script to run it, etc I want to use DTS and have it see an XML file the same way it sees any other data source. It should inspect the XML file, give me a list of tables/types and let me map them to existing data.
Data Validation Rules
- I won't elaborate here, but the data I've been working with is of unknown quality. As such I would love a way to easily validate that data. SQL has some easy validation rules (data type, nullability, joins), but I'd love more, like the ability to say that a date must be within a valid range, or that a number can't be negative, etc. I also want to be able to run these validation tests on data that already exists in a DB, in addition to running it as part of the data import process.
Unit Test Support
- Given the above point on unknown data quality, I would love a way to stub out unit tests to test for data integrity, as well as to test results of queries/views/stored procs against dummy data. You could even have code coverage tools.
- Rename - I would love a Rename refactoring for variables, table/view data types and any other DB objects. The refactoring would automatically check and update any dependencies, comments, etc just like VS 2005.
- Extract to Stored Procedure/Function - How many times have you seen a 10+ page query that is completely procedural and poorly organized? You'll find 1/2 of the code copy/pasted into about five other stored procs too. Having the ability to refactor and componentize queries (where possible) would rock.
- Extract to Lookup Table - Take a table Customer with a varchar(50) column named State with values like "Alabama, Arkansas,...". Extract the unique values to a Lookup table with the structure like (State_PK, State_text) and have a wizard walk you through setting up the state lookup table name, PK name and type, and take the extracted values/datatypes (the state names) automatically add it to the lookup table in a column with the same name/data type. The original table would of course remove the State varchar column and instead add a foreign key relationship based on State_PK.
- Convert Columns to Rows or Convert Rows to Columns- There may be a time when you have data of an unknown structure that you use a Row based table structure. A row based structure doesn't define the specific columns, but instead has only a couple of columns like Key and Value. There are times when you need to convert an already existing DB design between these. Although more rare, it's definitely a pain, and something that you could easily automate.
- Promote Local to Parameter - This would be nice to have, in particular if it was integrated with the Client Code Dependency analysis feature (below).
- Client Code Dependency Analysis - Assuming an ASP.NET example, I want to know which pages will break if I change my stored procedures, and if it's possible, automatically fix the changes. Let's say I rename a stored proc, the Web page that calls that stored proc won't work anymore. This would automatically update the reference to the correct name.
- Normalization Analysis/Automation - Show me where my DB isn't normalized and refactor the database (ex remove calculated columns). It doesn't have to hit some uber normalization like 6NF, but how about to 3NF? The key, the whole key, and nothing but the key, so help me Codd!
- SQL Server already has a Best Practices analyzer, which is pretty good, but I want it integrated into Query Analyzer. SQL Server BPA includes rules for maintenance and design issues, and I believe query "no-no's" like cursors which is also great. <aside>: Cursors are the biggest, easily replaced perf hitters I know of, and I've always wanted to start "nodatabasecursors.com" with examples of cursors and why people think they need them and how they can be replaced with SQL.</aside>
Source Control for Stored Procs, Views, Triggers, Functions
- Nuff said
My dream scenario: I would be able to import data using XML Bulk Insert through a wizard, run some validation code pre/post load, create a stored proc and refactor it automatically, generate unit test stubs for the stored proc, validate via FX Cop and then check in my code to some source control mechanism.
For the DB programmers out there, what other dev tool IDE features do you miss when writing SQL code? Are there any cool tools/add-ins that have the features I listed above that I might not know about?