January 13, 2010 at 5:58 pm
hi there. im new to the forums so please inform me if this is the wrong place to post the following:
i have a table called say, 'people' in ms sql server 2005 with the following fields:
person_id (int), isMale (bool), age (int), height (real), weight (real), numDataPoints (int), .......
the table currently has 1000 records partially filled with data. numDataPoints stores a point system based value depending on how much data each person has for that row. the more data filled in for that person during data entry stage, the higher the value.
what i would like to do is to simulate say, 2000 more records to make the total count 3000. i would like to fill in this 'dummy data' in such a way that the averages, ranges, standard deviations, etc are the same before the simulation and after.
i suppose i would do the following:
- first randomize a value for numDataPoints based on existing data for that field. I can do something like take the average and range on which to base a randomization.
- person_id - for this variable, i would have write a function to pick a new unique id
- i would find the average age, age range, standard deviation maybe?
i hope im clear with what im looking for. am i wrong to go about doing this using sql stored procedures? is there software out there that automates this? spss ? i appreciate any help! paul.
January 13, 2010 at 10:44 pm
info 68917 (1/13/2010)
...what i would like to do is to simulate say, 2000 more records to make the total count 3000. i would like to fill in this 'dummy data' in such a way that the averages, ranges, standard deviations, etc are the same before the simulation and after.
...
Why not just take the 1000 records you have and duplicate it twice? This would allow for the statistical data to be the same for the before and after simulation runs.
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
January 13, 2010 at 11:24 pm
great idea.. except they have to seem like unique individuals.. simply duplicating records isn't enough. i have decided to duplicate and muck around with the values to make them seem different but this is a quick and dirty solution..
i guess my point is.. what technologies would be the best solution... ive been coding this stuff in sql and have my method is too RAW - maybe there are tools out there to automate this.
January 13, 2010 at 11:33 pm
RedGate has a tool called Data Generator - that may be of help.
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
January 13, 2010 at 11:42 pm
excellent.. thanks!
January 13, 2010 at 11:46 pm
You're welcome
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
Viewing 6 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply