Rails spinoffs - Sep 2006 - Problems with large amounts of data using prototype's Ajax.Updater class

I have a asp page that generates about 1200+ records of data. When I
take the results and insert them into my container div using the
Ajax.Updater class the browser crashes or if it does load the browser
operates *VERY* slow. I tried to chunk the data out into smaller
amounts but I still get the same results. Is this fixable or am I just
trying to do too much with AJAX?


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs
-~----------~----~----~----~------~----~------~--~---

I have a asp page that generates about 1200+ records of data. When I
take the results and insert them into my container div using the
Ajax.Updater class the browser crashes or if it does load the browser
operates *VERY* slow. I tried to chunk the data out into smaller
amounts but I still get the same results. Is this fixable or am I just
trying to do too much with AJAX?


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs
-~----------~----~----~----~------~----~------~--~---

It''s probably not that your trying to do too much with AJAX,
there''s just
probably a more efficient way to implement your particular
page/functionality than what you''re doing. Having said that, we really
like
live demos here... put the problem page up somewhere and send us the link so
we can dissect it for you.

On 9/6/06, ringerce <cringer-FHbQlL4gZshBDgjK7y7TUQ@public.gmane.org>
wrote:
>
>
> I have a asp page that generates about 1200+ records of data. When I
> take the results and insert them into my container div using the
> Ajax.Updater class the browser crashes or if it does load the browser
> operates *VERY* slow. I tried to chunk the data out into smaller
> amounts but I still get the same results. Is this fixable or am I just
> trying to do too much with AJAX?
>
>
> >
>

-- 
Ryan Gahl
Application Development Consultant
Athena Group, Inc.
Inquire: 1-888-919-8700 x2903
Blog: http://www.someElement.com


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs
-~----------~----~----~----~------~----~------~--~---

Well, the problem with that is that the application that I''m building
is mean''t for internal use only and it would be against my companies
policies to make it public. I''m really not doing anything outside of
the box here. My page does return a lot of html and I think the browser
just runs out of memory. It seems to crash when the DOM parser comes
into play.

Here is the function I wrote though to update my container. Maybe
you''ll see something unusual:

function ajaxUpdater(mForm,mUrl,mMethod,outputObj) {
	var params = '''';
	if(mForm) {
	params=customSerializer(mForm)+''&time=''+currentTime();
	} else { params = ''?time='' + currentTime(); }
	var ajax = new Ajax.Updater({
		success: outputObj},mUrl,{
		evalScripts:true, method: mMethod, parameters: params,
		onFailure: function(response) {
			reportError(response.responseText);
		}
	});
}

------------------------------------------------------------------------------------------
and I call it like so
------------------------------------------------------------------------------------------
ajaxUpdater('''',''viewer.asp'',''get'',''contentPanel'');
------------------------------------------------------------------------------------------
Pretty simple. Lemme know if you spot anything unusual.


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs
-~----------~----~----~----~------~----~------~--~---

Hmm, nothing strikes me. A 1200 row table shouldn''t be that big of a
deal
(couple of seconds to render maybe depending on table and cell data sizes,
but then it should all be good), unless maybe you''re attaching event
handlers to every row/cell/element or something like that upon render, in
which case I''d say employ a lazy loading technique to delay event
wiring
until the last possible moment, depending on the event. Investigate how many
actual objects you''re creating in memory during each stage of the
request/render sequence (remember event handlers take memory too)... try to
eliminate anything unnecessary. And also, is this happening the first time
the data is loaded or only during subsequent requests (in which case
I''d say
you have a memory leak to sniff out)? So many things to think about :-)

Of course 1200 rows x 30+ columns x 1kb data per cell would be big... lol.
You may just have to resort to paging your data (although I''m assuming
this
option has already been discussed and for whatever reason there is a
requirement that all 1200+ rows be served at a time).

It''s just hard to say what your problem is without seeing something. I
mean,
it could even be your machine or network for all we know.

Anyway, sorry I have nothing more profound to say than that...

On 9/6/06, ringerce <cringer-FHbQlL4gZshBDgjK7y7TUQ@public.gmane.org>
wrote:
>
>
> Well, the problem with that is that the application that I''m
building
> is mean''t for internal use only and it would be against my
companies
> policies to make it public. I''m really not doing anything outside
of
> the box here. My page does return a lot of html and I think the browser
> just runs out of memory. It seems to crash when the DOM parser comes
> into play.
>
> Here is the function I wrote though to update my container. Maybe
> you''ll see something unusual:
>
> function ajaxUpdater(mForm,mUrl,mMethod,outputObj) {
>         var params = '''';
>         if(mForm) {
>        
params=customSerializer(mForm)+''&time=''+currentTime();
>         } else { params = ''?time='' + currentTime(); }
>         var ajax = new Ajax.Updater({
>                 success: outputObj},mUrl,{
>                 evalScripts:true, method: mMethod, parameters: params,
>                 onFailure: function(response) {
>                         reportError(response.responseText);
>                 }
>         });
> }
>
>
>
------------------------------------------------------------------------------------------
> and I call it like so
>
>
------------------------------------------------------------------------------------------
>
ajaxUpdater('''',''viewer.asp'',''get'',''contentPanel'');
>
>
------------------------------------------------------------------------------------------
> Pretty simple. Lemme know if you spot anything unusual.
>
>
> >
>

-- 
Ryan Gahl
Application Development Consultant
Athena Group, Inc.
Inquire: 1-888-919-8700 x2903
Blog: http://www.someElement.com


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs
-~----------~----~----~----~------~----~------~--~---

ringerce wrote:
>Well, the problem with that is that the application that I''m
building
>is mean''t for internal use only and it would be against my
companies
>policies to make it public. I''m really not doing anything outside
of
>the box here. My page does return a lot of html and I think the browser
>just runs out of memory. It seems to crash when the DOM parser comes
>into play.
>
>  
>
So, have you tried just loading up the data in its own page? Rendering a 
big table taxes the brower - especially IE. That would give you 
something to compare it with. Also, look at your table code - it should 
render faster with explicit column widths. I had to troubleshoot a 
similar problem recently where it turns out a customer was sending 5MB 
of html tables over the wire and it was freezing up the browser for 
minutes at a time. That code also happened to have inline event handlers 
(onmouseover etc) in it. (I actually tried removing those and using a 
single event handler and event.target filtering to get the same effect, 
bit counter-intuitively it didnt really have much effect on overall 
performance.)
With large amounts of data, every little inefficiency is magnified - so 
check all your tags are closed properly, remove unnecessary markup, 
validate etc.
>	var ajax = new Ajax.Updater({
>		success: outputObj},mUrl,{
>		evalScripts:true, method: mMethod, parameters: params,
>		onFailure: function(response) {
>			reportError(response.responseText);
>		}
>  
>
Are you actually sending script along with this data? If not, set 
evalScripts: false and you should see a big speed up.

Oh, and check with a fresh browser each time, and keep any eye on your 
task manager / CPU - if you are leaking memory anywhere, your testing 
can be skewed by previous versions of the page having crippled your 
browser.

Another thing: debugging. If you are writing debug statements to the 
same DOM you are trying to measure performance with its going to slow 
down. Even console.log in Firebug drags appreciably.

I''ve also found that render time and network latency arent as seperate 
as youd like them to be. If latency increases, overall performance can 
decrease in a non-linear way.

hth
Sam (-i-am)


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs
-~----------~----~----~----~------~----~------~--~---

Thanks for the input fella''s. Mucho appreciation. To answer some of
your question:

- This is happening when the page first loads and not on subsequent
page loads.
- I do have some onClick even handlers assigned to href''s in my markup
but right now I haven''t attached anything to them as you can see below.
I checked IE''s memory usage at runtime it''s at a whopping
150mb ..
yikes ...
- I do have some js that''s need to be evaluated on that page. I suppose
I could break up the page into seperate AJAX calls and just load the
data table seperately since it doesn''t contain any js that needs to be
evaluated.
- Also, I can replicate the problem in all browsers pretty much. IE,
Firefox, safari, opera ...

Let me show you what the resulting HTML output would be. I''m not using
tables.
<div id="comment_1">
    <div class="tRowHeading bg1">
      <div class="float-Left">
      <span class="fs12b">beta1</span><br/>
      <span class="fcdgray">Approved:
</span><strong>No</strong><br/>
      <span class="fcdgray">Resolved:
</span><strong>Yes</strong>
    </div>
    <div class="float-Left pt6">
      <span class="tCell_aligned">Screen: </span>
      <span class="fs12b">010000000010</span><br
clear="all"/>
      <span class="tCell_aligned">Assigned To: </span>
      <span class="fs12b">Justin Bailey</span>
    </div>
    <div class="dtStamp">
      <span class="fcdgray">Last Updated:
</span><strong>1/22/2006
7:41:00 AM</strong>
    </div>
  </div>

  <div class="tRow bg1">
    <h1>Comment</h1>
    <div class="tCell"><a onClick=""
href="#"
class="btnAddComment"><span>Add
Comment</span></a></div>
    <div class="tCell"><strong>Jon
Doe</strong><br/>3/10/2006, 12:09:52
PM</div>
    <div class="tCell c3"><p>wish that at the end of the
session you
had an option to continue with the next lesson or go back to the table
of contents. <BR>Might want to consider having a check box next to the
lessoon on the TOC when you have completed that lesson, so its easier
to click on the next one that you want. n</p></div>
    <br clear="all" />

    <div class="hr"><hr /></div>
    <h1>Interal Actions</h1>
    <div class="tCell"><a onClick=""
href="#"
class="btnAddComment"><span>Add
Comment</span></a></div>
    <div class="tCell"><strong>Jon
Doe</strong><br/>3/10/2006, 12:09:52
PM</div>
    <div class="tCell c3"><p>Menu will be in place for
Beta 3 release
and should resolve this issue.</p></div>
    <br clear="all" />
  </div>
</div>

---------------------------------------------------------------------
So that''s one row of data. All that x1200 is a lot of markup to
display. I''ve already implemented the paging system and I think
I''m
going to create an "EXPORT ALL RECORDS TO EXCEL" or something like
that
so they can have what they want.


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs
-~----------~----~----~----~------~----~------~--~---

Oh, also one more thing. I did try loading the page outside of my AJAX
framework by itself. The page still performs slowly on it''s own and
still crashes sometimes. If I remove my stylesheet then it loads fine
and operates smoothly.

Just food for thought.


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs
-~----------~----~----~----~------~----~------~--~---

Well then, looks like you''ve isolated the bottleneck. Next step:
optimize
your stylesheet :-)

Also, just a nit picky thing (take with a grain of salt), but you really
shouldn''t be sending A tags (or any tags for that matter) with
onClick="". I
(and most people in this list) recommend getting into the habit of attaching
behavior outside of the markup. But even moreso, rendering empty attributes
is just bad form (offers no symantic meaning). I''m not saying that
necessarily has anything to do with your performance issues (especially if
you say that removing the stylesheet solves that)... just something to think
about.

On 9/7/06, ringerce <cringer-FHbQlL4gZshBDgjK7y7TUQ@public.gmane.org>
wrote:
>
>
> Oh, also one more thing. I did try loading the page outside of my AJAX
> framework by itself. The page still performs slowly on it''s own
and
> still crashes sometimes. If I remove my stylesheet then it loads fine
> and operates smoothly.
>
> Just food for thought.
>
>
> >
>

-- 
Ryan Gahl
Application Development Consultant
Athena Group, Inc.
Inquire: 1-888-919-8700 x2903
Blog: http://www.someElement.com


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs
-~----------~----~----~----~------~----~------~--~---

ringerce wrote:
>Oh, also one more thing. I did try loading the page outside of my AJAX
>framework by itself. The page still performs slowly on it''s own and
>still crashes sometimes. If I remove my stylesheet then it loads fine
>and operates smoothly.
>  
>
I''ve had this kind of trouble before with a floated layout. I think 
floats are hard work to render. Try it (just for kicks if the idea is 
offensive) in a table and see if it improves. I''m a css layout fan, but
this is part of the trade-off - you leave a lot for the browser to 
calculate as it draws the page.

Sam (-i-am)


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs
-~----------~----~----~----~------~----~------~--~---

Plus tables are great for tabular data. Not sure why people think any use of
tables is offensive. Tables for general layout, yes it''s wrong, for
simple
tabular data display, not a problem and usually the correct choice over any
alternative.

On 9/7/06, Sam Foster <sam-WNdku5qRq3FWk0Htik3J/w@public.gmane.org>
wrote:
>
>
> ringerce wrote:
>
> >Oh, also one more thing. I did try loading the page outside of my AJAX
> >framework by itself. The page still performs slowly on it''s
own and
> >still crashes sometimes. If I remove my stylesheet then it loads fine
> >and operates smoothly.
> >
> >
> I''ve had this kind of trouble before with a floated layout. I
think
> floats are hard work to render. Try it (just for kicks if the idea is
> offensive) in a table and see if it improves. I''m a css layout
fan, but
> this is part of the trade-off - you leave a lot for the browser to
> calculate as it draws the page.
>
> Sam (-i-am)
>
>
> >
>

-- 
Ryan Gahl
Application Development Consultant
Athena Group, Inc.
Inquire: 1-888-919-8700 x2903
Blog: http://www.someElement.com


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs
-~----------~----~----~----~------~----~------~--~---

The data that I''m outputting has some fancy styling so with tables
I''m
just creating an unmangable mess for myself. Anyway, I''ll try and
create a simple table version of my layout and see what happens. Be
back later. I cringe to type "<table>".


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Spinoffs" group.
To post to this group, send email to
rubyonrails-spinoffs-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-spinoffs-unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-spinoffs
-~----------~----~----~----~------~----~------~--~---