What Do We Need in SQL Server?

  • Having the ability to "variableize" a few commands like BULK INSERT and OPENQUERY would be nice.

    ...something like:

    BULK INSERT dbo.MyTable

    from @FileSpec

    with ...

    ...and for open query...

    create proc myproc @key int

    as

    declare @sql varchar(max) = 'select columns from table where key = ' + cast(@key as varchar(10))

    select columns from openquery(OracleLinkedServer, @sql);

    --Paul Hunter

  • Allow database locations to be stored in Active Directory. I want to my connection string to point to "AdventureWorks", and AD should be consulted to determine which server/database to connect to. You then end up with a single source of truth as to where an application should be pointed to, and it's a single point of configuration change when a database moves.

    I'm sure there's a Connect item for this, I just can't find it right now!

  • When I used to work on Burroughs / Unisys mainframes , the proprietary DbMS (DMSII) has the ability to perform an archival rollback , i.e. you roll the database back in time through as many logs as needed rather than restore from a backup and roll forward - this was often very useful in the event of overnight batch failure where a rollback you take say 20-30 mins to rollback 5-6 hours of batch as opposed to a restore from backup and roll forward which could take several hours .

    A feature like that could be very useful in SQL server

  • Apologies if this was already mentioned...something akin to Oracle's LAG function that can optionally be accumulated in a variable. This would obviate the need for CLR and triangluar joins for running totals

Viewing 4 posts - 31 through 33 (of 33 total)

You must be logged in to reply to this topic. Login to reply