Thanks all for the insight!


>>>>

>>>>NewEmpID=5678
>>>>select employee
>>>>scan for employee.empid=1234
>>>>  currec=recno()
>>>>  insert into employee(empid,task);
>>>>     values (NewEmpId,employee.task)
>>>>  goto currec
>>>>endscan
>>>>

>>>
>>>Unless you are doing some other processing within the SCAN...
>>>I think the following approach gives an equivalent result. How about
>>>
>>>

>>>SELECT 5678 AS empid,task FROM employee;
>>>  WHERE empid=1234;
>>>  INTO CURSOR curTemp
>>>
>>>SELECT employee
>>>APPEND FROM DBF('curTemp')
>>>

>>
>>Is there a reason your method would be better?
>
>Cause I thought of it???? <g>
>
>FW(Little)IW
>
>My impression is that taking things out of loops and processing the entire file in one command is usually faster and cleaner. For example
>
>

REPLACE ALL myField WITH somevalue FOR condition

>rather than
>

SCAN FOR condition
>     REPLACE myField WITH somevalue
>ENDSCAN
>

>
>By taking all the extra record changing out of the equation I think my method will be significantly faster. Of course, significantly faster is a relative term and a large percentage improvement may not be meaningful.
>
>I used big (200000 records) tables with lots of replacements. My tests showed a drop in processing time (5.14 seconds vs 3.83 seconds).
>
>My .02 cents...............Rich
>
>Oh, something else. I don't have VFP 7 or SQL Server. Could one of the real gurus tell me if
>
>

INSERT INTO employee ;
>       (SELECT 5678 AS empid,task FROM employee e2;
>         WHERE e2.empid=1234)
>

>
>be a legal statement in either???