After running this in SQL 2008 I've discovered that the stored proc doesn't pick up the last run's data.

Changing the INNER JOIN's condition from INNER JOIN dbo.sqlio_import impfulltest ON imp.rowid + 20 = impfulltest.rowid AND impfulltest.resulttext = " to INNER JOIN dbo.sqlio_import impfulltest ON imp.rowid + 19 = impfulltest.rowid fixes the problem. (A non-coding workaround is to append a blank record with the maximum ParameterRowId to dbo.sqlio_import before running the proc.)

The script.

Okay, The script was modified by SQLGuyChuck last December 19, 2009 (I'm assuming Chuck Lathrope). In it he gives an overhaul to the script:

• Using 6 testfiles instead of 1.
• redirecting output to hardcoded h: (as opposed to the method described in the video).
• renaming SQLIO_1pass.bat to SANTest.bat
• Where does "timeout /T 10" come from. It doesn't seem to be an SQLIO parameter or a dos command.

These improvements are fine maybe, but there's no, text or comments that explain the purpose of the changes. The video and everything around it become out of sync. The mismatch, among other things, makes me (and probably others) a little hesitant and wary about the info. I'm not sure why 6 files are better than one. What do the test results mean?

Personally I'm looking into the history of this article and plan to use the original script.