Comments on: Sample Standard Deviation (and SQL) http://blog.todmeansfox.com/2006/12/12/sample-standard-deviation-and-sql/ Business Intelligence, Data Warehousing, SQL, Visual FoxPro. Thu, 20 Nov 2008 22:50:06 +0000 http://wordpress.org/?v=2.2 By: Anders Altberg http://blog.todmeansfox.com/2006/12/12/sample-standard-deviation-and-sql/#comment-653 Anders Altberg Sat, 05 Jan 2008 19:32:18 +0000 http://blog.todmeansfox.com/2006/12/12/sample-standard-deviation-and-sql/#comment-653 Hi Todd, The Std (standard deviation) function can be done entirely in VFP SQL. Can be combined with GROUP BY which can save a lot of code. Assumin a numerical column 'score' in a table 'Scores': SELECT SQRT(AVG(score^2)-AVG(score)^2) FROM Scores A Samples STD can also be calculated: SELECT SQRT(COUNT(*)*(AVG(score^2)-AVG(score)^2)/(COUNT(*)-1+0.000001)) FROM Scrores or a Variance SELECT COUNT(*)*(AVG(score^2)-AVG(score)^2)/(COUNT(*)-1+0.000001) FROM Scores The +0.000001 getsa around a possible divide by zero error. Cheers Anders Anders Altberg Pragam Hi Todd,
The Std (standard deviation) function can be done entirely in VFP SQL.
Can be combined with GROUP BY which can save a lot of code.
Assumin a numerical column ’score’ in a table ‘Scores’:
SELECT SQRT(AVG(score^2)-AVG(score)^2) FROM Scores
A Samples STD can also be calculated:
SELECT SQRT(COUNT(*)*(AVG(score^2)-AVG(score)^2)/(COUNT(*)-1+0.000001)) FROM Scrores
or a Variance
SELECT COUNT(*)*(AVG(score^2)-AVG(score)^2)/(COUNT(*)-1+0.000001) FROM Scores

The +0.000001 getsa around a possible divide by zero error.

Cheers
Anders

Anders Altberg
Pragam

]]>