Results 1 to 7 of 7
  1. #1
    Join Date
    Jan 2004
    Posts
    26

    Unanswered: how to design a thorough test plan?

    I am going to handle a test to a list of querys to find their efficency(spending of time)

    here is my test plan:

    there are query ABC..., and insert all querys into a table called querytbl;
    open a cursor for all records from querytbl;
    fetch next query from cursor;
    while @@fetchstatus = 0
    begin
    exec query for 3 times and calculate average spending of time;
    fetch next query from cursor;
    end
    ...

    Is there any better test plan?(just test spending of time)
    or test tools?
    ......

  2. #2
    Join Date
    Nov 2003
    Location
    Christchurch, New Zealand
    Posts
    1,618
    One of the things I think you would want to include is the changing of parameter data (if applicable).

    by that I mean that...

    select * from tblMyTest where MyID = 12345

    might return a lot faster then

    select * from tblMyTest where MyID = 54321

    depending on how the tables have been constructed.

    You probably want to test with different levels of data as well eg, 10000 record, 1000000 records etc.

    What exactly are your trying to prove by your testing? Performance obviously, but are you also stress testing, load testing and durability testing, all of which are performance related.

    HTH.

  3. #3
    Join Date
    Jan 2004
    Posts
    26
    The main purpose of the test plan is to compare perfomance of the same querys to different databases which have same data but different Logical/phsical structure, or to compare performance of different versions of the same query to same database.
    -------may call it "test different structure's performance"?
    we do that because we want to get a general contractive performance report of all querys or versions when we want make some change to databases or querys, that'll help us to decide whether to apply the change.
    ......

  4. #4
    Join Date
    Nov 2003
    Location
    Christchurch, New Zealand
    Posts
    1,618
    Okie, well in that case one of the things you probably want to include in your testing is how the query performs when other activities are taking place on the database tables that the query is referencing.

    You may find that despite the fact that 70% or the time the 3 seconds query is faster, 30% of the time the query take 10 seconds longer because of the locking that is involved in the query.

  5. #5
    Join Date
    Jan 2004
    Posts
    26
    thanks! Actually All querys is executed in sequence in a batch,and there is only one batch running,we will stop other clients also,so I think In that case wonnt occur a lock.
    one thing I am not sure is that whether a query will run faster or later if the query was run in different order in sequence?
    ......

  6. #6
    Join Date
    Nov 2003
    Location
    Christchurch, New Zealand
    Posts
    1,618
    I can't think of any reason why it would,.... but you might want to try it just to make sure....

  7. #7
    Join Date
    Jan 2004
    Posts
    26
    thanks for advises!
    Last edited by cinrain; 02-25-04 at 11:37.
    ......

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •