Login     Register

        Contact Us     Search

XLeratorDB/statistics Documentation

SQL Server F.TEST function


 
Updated: 6 August 2010

Use FTEST_q to return the result of an F-test. An F-test returns the two-tailed probability that variance in dataset 1 and dataset 2 are not significantly different. Use this function to determine whether two samples have different variances.
 FTEST
Syntax
SELECT [wctStatistics].[wct].[FTEST_q] (
   <@Values1_RangeQuery, nvarchar(4000),>
 ,<@Values2_RangeQuery, nvarchar(4000),>)
Arguments
@Values1_RangeQuery
the select statement, as text, used to determine the first set of values to be used in the FTEST_q calculation.
@Values2_RangeQuery
the select statement, as text, used to determine the second set of values to be used in the FTEST_q calculation.
Return Types
float
Remarks
·         If the number of rows in dataset 1 or dataset 2 is less than 2, FTEST_q will return an error.
·         If the variance of dataset 1 or the variance of dataset 2 is zero, FTEST_q will return a divide by zero error.
·         For simple queries, consider using FTEST.
·         No GROUP BY is required for this function even though it produces aggregated results.
Examples
CREATE TABLE #data1(
      [num] [float] NOT NULL
)
INSERT INTO #data1 VALUES (6)
INSERT INTO #data1 VALUES (7)
INSERT INTO #data1 VALUES (9)
INSERT INTO #data1 VALUES (15)
INSERT INTO #data1 VALUES (21)
CREATE TABLE #data2(
      [num] [float] NOT NULL
)
INSERT INTO #data2 VALUES (20)
INSERT INTO #data2 VALUES (28)
INSERT INTO #data2 VALUES (31)
INSERT INTO #data2 VALUES (38)
INSERT INTO #data2 VALUES (40)
Select wct.FTEST_q('Select num from #data1', 'Select num from #data2')


This produces the following result
----------------------
0.648317846786174
 
(1 row(s) affected)


Copyright 2008-2013 WestClinTech LLC         Privacy Policy        Terms of Service