>I am modifying an application that has many instances of multiple
>REPLACE commands to update tables, such as
>
>
>replace field1 with "value1"
>replace field2 with "value2"
>replace field3 with "value3"
>
>
>I want to speed this application up. So I considered:
>
>UPDATE MyTable;
> SET Field1 = "Value1";
>   Field2 = "Value2"
>   Field3 = "Value3"
>
>
>Since REPLACE starts at the first record, the 3 separate REPLACE commands
>therefore make 3 passes through the table, while the SQL Update goes through
>the table only once.
>
>However, the Hackers Guide says the SQL Update is slower than REPLACE. Isnt
>REPLACE 'old school' and SQL 'optimized'?
>
>What are the Pros and Cons of each method?
>
>Thanks

First of all, you'd be better off with

REPLACE ;
  field1 WITH "value1", ;
  field2 WITH "value2", ;
  field3 WITH "value3", ...
  IN TableAlias

Second, the default scope of REPLACE is NEXT 1, so it acts on the record at the record pointer position - no pointer manipulation or optimization required, should be faster than UPDATE - SQL. Get used to using the IN clause, which prevents the xBASE "feature" of REPLACE not working if the table to be updated is not the current work area, and you're at EOF() in the current work area.

UPDATE - SQL is more "modern" in that it is SQL-style, the only type of update that SQL backends support, so it makes some sense to get used to that type of syntax. It is also a much more powerful command than REPLACE if you want to work with multiple records, as you can use SQL syntax and set logic to select records for updating.

Each command has its place, IMHO.