>>>>Using macro substitution inside a query is a bad idea anyway because it forces SQL to evaluate the macro for every row, and slows the query dramatically. HTH
>>>
>>>Erik, are you sure about this? If yes, then try some quick test.
>>>Actually, this is UDF which might be evaluated for each row.
>>
>>No, I am not sure, Ed, it's just what I have always been told. (This was also discussed in one of the DevCon sessions... I'll have to look at my notes to remember which one.)
>>
>>I did one quick test though...
>>
>>lfilter = "lname = 'Moore'"
>>before1 = seconds()
>>select * from campus into cursor cursor1 where &lfilter
>>after1 = seconds()
>>elapsed1 = after1 - before1
>>wait window "Query1: " + str(elapsed1)
>>
>>Wait window returned "Query1: 23"
>>
>>I then exited VFP, returned and changed the code to:
>>
>>before2 = seconds()
>>lquery = 'Select * from campus into cursor cursor2 where lname = "Moore"'
>>&lquery
>>after2 = seconds()
>>elapsed2 = after2 - before2
>>wait window "Query2: " + str(elapsed2)
>>
>>Wait window returned "Query1: 11"
>>
>>I realize this isn't the most dependable of tests, but nothing better came immediately to mind. Have you tested for the performance difference? I am interested to hear the results.
>
>As I understand Foxpro will first evaluate macros, compile whole command and run query, so macro will be evaluated once. My tests don't show time difference. I will appreciate replies from other UT members on this topic.

Ed- on review of my DevCon notes, I believe you're right. I think that I was mentally combining the results of two facts:

1. SELECT statements will evaluate UDFs for every line.
2. Macros slow down a scan loop because VFP re-evaluates them