When you're talking about "read the data", what type of object would you be
reading from? If it's a Result generated by one of the tags, then any
significant performance hit will already have been taken. In other words, the
*potentially* slow and memory intensive part of the SQL tags is in
"snapshotting" a ResultSet as a Result. This is really not too different from
using something like a disconnected RowSet. If you're talking about reading
directly from a ResultSet, then that implies the type of JSP that nightmares are
made of ;-).

As an aside, I've done my own micro benchmarking on JSTL's Result vs. BeanUtil's
RowSetDynaClass (another way to snapshot a ResultSet). It turned out that Result
objects were generated more quickly. I wouldn't necessarily categorize it as a
"significant difference", however, and I haven't compared memory consumption.

Quoting Henri Yandell <[EMAIL PROTECTED]>:

> 
> Just guessing, but I doubt it will be faster performance-wise, just faster
> to develop. It lets you treat JSP as a scripting language and remove the
> compile/deploy stage.
> 
> Hen
> 
> On Thu, 5 Feb 2004, Riaan Oberholzer wrote:
> 
> > Is the SQL tags (in JSTL) performance acceptable?
> >
> > I'm populating tables and was wondering if it would be
> > better to read the data in "normal" Java code and then
> > populate the table with
> >
> > <td><%= row.getItem1() %></td>
> > <td><%= row.getItem2() %></td>
> >
> > etc.
> >
> > I'm not too worried about trying to keep java code out
> > of jsp's, I really want the fastest solution if it can
> > make a significant difference.

-- 
Kris Schneider <mailto:[EMAIL PROTECTED]>
D.O.Tech       <http://www.dotech.com/>

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to