It's unknown how VFP handles such conditions but most likely it creates a cursor with subquery and then matches each record from the main table to it. The following command should be faster

UPDATE Invoices ;
		SET InvoiceId = "ABC123" ;
	FROM Invoices 
		JOIN Invoices2 ON Invoices2.RecordId = Invoices.RecordId

>I have the following command:
>
>

>UPDATE Invoices;
>  SET InvoiceId = "ABC123";
>  WHERE RecordId IN;
>    (SELECT RecordId
>     FROM Invoices2)
>
>

>
>
>
>1) Does VFP first create a subset of Invoices2.RecordId and then use that
>to drive the update Invoices? Or is each record in Invoices2 evaulated
>for every record in Invoices?
>
>2) Is there a more efficient way to do this?
>