Re: ParsedStatement Behavior

On Tue, 28 Nov 2006, alfranio correia junior wrote:

> Please, if there is any archive on this let me know.
> I was wondering if the reasons to have an specific protocol in order to 
> prepare statements are:
>
> 1 - Previous versions of the driver used PREPARE and EXECUTE to implement 
> server-prepared statements. This is supported on all server versions 
> beginning with 7.3, but produced application-visible changes in query 
> results, such as missing ResultSet metadata and row update counts.
>
> 2 - The need to transfer strange data types such as circle, point.
>
> 3 - To Improve performance as binary data type is usually shorter and avoid 
> converting from text to binary and vice-versa.
>
> Disregarding such reasons, might I use "prepare name statement" ?

There's nothing stopping you from doing so.  The only potential problem is 
that I think one of the older server versions does not take bind 
parameters passed to execute, but a recent release should avoid that.  The 
other question is why?  The only advantage that I can see is that you have 
better control of the lifespan of prepared statements.

Kris Jurka