>Hi,
>OK, it would take me several days to regenerate the test environment and redo the tests, but if I do this, I would require you to also justify all general observations that you make in the future, do you wish to go down that road.

I'm ready to back up what I say with evidence. I can't even be sure if I remember things correctly, so there is merit in redoing the tests, especially in different versions. In that way, if the code is wrong or the assumptions are wrong, someone should point out the flaw in the code or assumptions and we all learn.

>
>>>No, you do your own testing. It is not the SET EXACT but the SET ANSI that effects the SQL Selects, the SET EXACT is a logical preference, as if I do a SEEK I want an exact match not a generic match, if I want a generic I will SET EXACT OFF and set it back on after I am finished.
>>
>>As I tried to point out scientists have to prove their findings. That's a huge failing in computing.
>>
>>>
>>>I think that the VFP default being SET EXACT OFF is one of the the most confusing defaults I have ever seen.
>>>
>>>
>>>>>Hi,
>>>>>I was just stating my findings, it would appear that the "==" is evaluated on each iteration, whereas the SET EXACT & SET ANSI are not.
>>>>>
>>>>>Do some tests yourself, create a dbf with 15mil + records and a number of keys, create several keys based on integer, character & binary fields and see what your findings are. Conservatively it gives a 10% difference in timings, remember to reboot the workstation between tests to get reliable results.
>>>>
>>>>Well as Jim Booth points out == is not the same as SET EXACT, there is the issue of comparing apples to apples. It's your claim, demonstrate it. You've already written the code.
>>>>
>>>>>
>>>>>>>Hi,
>>>>>>>I have recently done some performance testing on large datasets in VFP9, as result we are now using SET EXACT ON & SET ANSI ON and not using "==" as it does have a perceivable impact on performance when processing large amounts of data.
>>>>>>
>>>>>>When a scientist wants to make such a claim in a journal they must provide evidence and practices which others can duplicate. Are we not doing computer science? IOW would you please prove that?
>>>>>>
>>>>>>>
>>>>>>>>Good Friday Morning to All,
>>>>>>>>
>>>>>>>>I ran into a problem using <> operator where the following happens:
>>>>>>>>

>>>>>>>>cTest1 = "This is a sample string"
>>>>>>>>cTest2 = "This is a sample string test"
>>>>>>>>? cTest2 <> cTest1  && returns .F.
>>>>>>>>

>>>>>>>>
>>>>>>>>Now I am thinking if I should do away with using <> operator because it always relies on SET EXACT ON and just use "==" in ALL my comparisons. What do you think?
>>>>>>>>
>>>>>>>>Thank you.