Thanks - this stuff will help - but it seems you're saying - if it the statement is good - then the only way we would know would be if an actual record were inserted into a table.

>>>What do you mean damage?
>>Thanks
>>
>>An insert or update implies stuff is to be written to a SQL table. The "wizard" has an interface wher the user can enter a SQL statement. If the SQL statement calls an SP or issues an INSERT from the client, there is data in the parameters that is written to the tables. I just want to assure that the statement is syntactically correct (fieldnames - statement structure) without actually writing to the table - and - without actually passing real data and inserting a record into the table.
>>
>>On the winzard for each statement is a "Test" button the user can click once the statement is in. If the SQL is an INSERT, I would like to assure the user that the statement is syntactically correct (when they press "Test") without actually writting anything to the table.
>
>If it's not syntactically correct (and semantically to an extent, i.e. has misspelled field names), sqlexec() won't do a thing and will return a -1 anyway.
>
>>With SELECT statements I can test without having to "gather" data - I can make select return a zero record cursor. I am looking for a "zero" action insert!
>
>I thought SqlPrepare() would actually compile and return some status, but it returns 1 for any gibberish you pass to it... but then, I got another bright idea: if your user clicks the TEST button, why don't you really send the command to the server, just wrapped into a begin transaction / rollback? No harm done that way, and you actually do get some status back.