PS: no, I don't think that a single line of code should be 8K long. So, if VFP had a function that took an unlimited length string and compiled the view, that would be fine, too. It's just that there isn't such a function, and doing it as a single line of code is what we have.

These are not typed, let me hasten to add. The view in xCase might have 20 joins (the xCase view designer is great once you become friends with it). We then, sometimes, add parts to the xCase view. Then we take that string and generate the view. There is no typing involved (thankfully).

So if it's easier to implement, having a function that takes a string and creates the view based on that string would be just fine. Something like

_vfp.createsqlview(<select string>,<view name>[,<dbc>,...<other parameters in CREATE VIEW command>).

It needs to be available at runtime as well as designtime (think: populating temporary dbc's).

Not that I expect you to implement it <s>, but just in case the VFPT is reading...

Hank

>Hank,
>
>Honestly, do you think a single line of code should be 8k?
>
>That is 100 lines of 80 column text!
>
>I really think this fits into the just because you can use a bajillion table join in VFP9, doesn't mean that you should category.
>
>>Absolutely agreed: it is useless having the wonderful new complex SQL capabilities, and being stuck at 8192 for SQL statement length!