views:

655

answers:

2

I have a bunch of tables like this:

Lookup_HealthCheckupRisks
------------
ID  Name
1   Anemia
2   Anorexic
3   Bulemic
4   Depression
...
122   Syphilis



PatientRisksOnCheckup
------------------
ID CheckupID RiskID
1  11        2
2  11        3
3  12        1
4  14        1
5  14        3
...

But I need a flattened version, like this:

PatientCheckup
------------------
CheckupID Risk_1 Risk_2 Risk_3 Risk_4 .. Risk_122
11        0      1      1      0         0
12        1      0      0      0         0
13        0      0      0      0         0
14        1      0      1      0         0

I'm clueless how to do this, the best I can think of is to write a temp table, define all 122 columns, and then do If Exists ( SELECT * FROM PatientRisksOnCheckup where RiskID=i and checkupID=j ) INSERT INTO PatientCheckup (1) WHERE CheckupID=j and iterate over i, j... >_<

Writing this query for just one table is doable not the best, but I've need to flatten data like this for another thirty tables of the same size. Er... suggestions please?

I am also curious to know if what I am doing is a common thing to do or not... ?

I am needing to denormalize/flatten the sql data for statistics software.

EDIT: It's in MS SQL Server.

+7  A: 

What you need is called a crosstab query.

If you're using Microsoft SQL Server, you can use the PIVOT operator to do it.

Other brands of RDBMS have varying support for this type of query. Worst case is you'll have to use dynamic SQL to hard-code very value from the lookup table into a join to your main table. This is not practical when you have 122 distinct values.

Also see SO questions tagged pivot or crosstab.

Bill Karwin
+1  A: 

Use PIVOT TABLE Here - Microsoft and here - tutorial.

You will need hovewer to specify all the columns. But you can use sp_executesql command to use dynamic SQL.

Lukasz Lysik