>Tom,
>
>If you have a newid.prg which contains the function make sure you don't have a line function NewID() in the code. Just comment it out and use LPARAMETERS if there are arguments. The NewID() function will work about a factor of 20 faster.
>
>Other than that you might check to make sure where the real slowdown is occuring. Opening and closing files is a fairly slow operation opening a few million files is just going to take a lot of time, especially if a network is involved.
>
>Are you doing string processing of the files? VFP string functions can be slow maybe calling a VC++ DLL would make it faster.
>
>You could temporarily take out the default value if you don't need it as a key for relational links, and just run a single REPLACE after the import to update the table.
>
>>I have a project where I need to load in a few million
>>text files to one of two tables. I am using a primary
>>key that is assigned via a stored procedure routine
>>similar to the one used in TasTrade (newid()).
>>
>>The first file contains 1.2 million records and after
>>22 hours it is only about 1/2 complete.
>>
>>Would it have been better to not set NEWID() as
>>the default value and used a different technique?

David, thanks for the response. The newid() function is truly
a stored procedure in the database container it is not a standalone
PRG.

I am curious as to how I could use a VC++ DLL since I do not
know C++. I am curious about the API section of the help file that
contains things such as _EdCopy and _EdPaste. I have no
idea how to use the _APIs. This is probably somewhat off
the original topic, but could you show me how these work or
do I need to have C++?