I am not a developer (not really a proper one) but always up for any excuses for self-development.

Sometimes inside a support team for any trustworthy software vendor, and that we presently make use of a helpdesk software program known as iSupport. - Its not necessarily a bad bit of package, and i am not confident that it's been setup badly (attempting to discover) however the greatest problem I face (standing on the leading line being an analyst who uses it 8hrs each day) is its lack of ability to simply search.

A brand new 'incident' will be. Client will report some errors inside a log, possibly mention another key phrases when explaining his signs and symptoms. Now we all know, I've most likely clarified an identical problem before (but cant recall the solution) or maybe more likely, another-analyst might have clarified exactly the same question before.

I must be capable of have one search box (think Google) that searches through EVERY incident which has have you been produced and return me ALL occurrences which contain that keyword.

Right now the search is extremely poor - You are able to make time to setup searches and specify which fields you need to explore, which values to filter by (possibly by an analyst or category, etc) but this needs time to work and generally, it returns poor results also it could have been simpler to track it lower yourself by hand.

All the data sits in underlying SQL Server tables (have asked for a subset from the data).

What im thinking, is developing a separate front-finish that's only a fundamental search box and thats it. This application points to any or all the appropriate fields within the SQL tables and take out the appropriate records right into a table. After I possess the ID for that incident, this will make it an easy job to drag out that incident in the iSupport front-end.

I believed like Google Desktop style application (shortcut key raises the search box).

Now thats just as much thinking as ive done. Searching for top tips on what to do next. I understand for example, that Google Desktop crawls and indexes all of your physical files in your machine. Would I must make a move similar for any database when i imagine there maybe a lot of records/fields/tables to look through.

TBH, whether it works, i am not that fussed (to start with) whether it takes some time to process the query, as lengthy because it returns relevant results. But ideally id want it to stop wasting time.

Ill let it rest at this for the time being. Where must i begin?

If .Internet is the factor, then Lucene.NET will fully trust SQL Server to provide you with that search feeling.

The StackOverflow websites utilize it, you are able to pay attention to the SO Podcast where Shaun/Joel bitch about why SQL Server full-text search sucks a lot.

I'd suggest this can be a great candidate for any web application - an asp.internet / jsf website. Which means that you are able to manage it in one machine, but all of your co-workers can take advantage from it with no deployment headache any time you give a new feature...

The incident database is mission critical (important to your relationship with clients), if you found me with this particular request I'd insist that you simply utilized the database via a user which had choose permissions towards the appropriate tables, and hardly anything else. This out of your perspective is really a positive thing too - let us you operate knowing you are not likely to cock anything up...

The SSMS Tool Pack (an add-directly into Management Studio) consists of an element to look Table, View or Database Data.

Take a look in to the Full-Text Search functionality of SQL Server.