Re: Support of a large table


"Dag Sunde" <me@xxxxxxxxxxxx> wrote in message
news:JBDPe.4675$qE.1082978@xxxxxxxxxxxxxxxxxx
> "Lee" <lee.p.baker@xxxxxxxxxxxxxx> wrote in message
> news:430efc15_1@xxxxxxxxxxxxxxxxxxxxxxxxx
> >I am trying to write a tool to extract the contents of an XML file and
> > display different parts within different controls of a vb app.
> >
> > One part of the XML file identifies information about processor cylces.
> > With the appropriate XSL style *** this information is displayed in IE
> > as
> > follows:
> >
> > Clock Cycle: 1 2 3 4 5 6
> >
> > Dispatch: A1
> > A2
> >
> > A2
> > A3
> > A3
> >
> > There are a few more rows but you get the idea. Basically the clock
cycle
> > count can go pretty high (30000)
> > Due to problems with IE/Netscape and memory issues with an XML file like
> > this, I thought about writing my own app using the SAX as it would be
less
> > demanding on memory.
> >
> > I basically wish to display this table in a suitable control. I dont
want
> > to just display only part of the table as it would mean re-parsing the
> > file
> > to get the next section and I guess storing the entire contents in
memory
> > could be rather hungry on memory.
> >
> > This table is necessary due to contractual reasons and on the odd
occasion
> > when the user may with to look at a particular part of it
> >
>
> How big is the complete (raw) xml-file?
>
> As long as you're going to write it in VB (Since IE fails), don't assume
> that
> Your VB app is going to suffer the same constraints.
>
> Try to open it in vb with DOM first (Instead of SAX). If the time it takes
> to
> load is not to long for your requirements, use DOM. Then you can keep the
> complete dataset in memory, and use XPath queries to extract "windows" or
> sub-sets of the data as the user request it.
>
> Then you can for example chop the datasets into chunks (visually).
> Ie. cycle 1-5000, 5001-10000... with a tab-***, buttons, a combo-box...
> etc.
>
> --
> Dag.
>
>
>
>

Thanks for the suggestions,

The raw XML file which I've tested is 7Mb with a cycle count of 14000 but I
believe there will be files up to 16Mb.

The 7Mb file is parsed using the SAX in about 10 secs and actually seems
fine. There is no processing on my part at this stage but first impressions
are ok.

IE doesn't exactly fail but I believe the DOM builds a tree and, as you
said, stores the tree in memory, the memory usage on my PC for the 7Mb
file goes from 210Mb to about 700Mb to open this single file. Obviously
trying to open a similarly sized file in another window is pretty futile. I
assume the same problem will occur with my app if I use DOM.

Just a note: I had a play with a flex grid and managed to populate 30000
colums with random text in about 15 secs and 40000 in 30 seconds but once
loaded seems fine. I may have a further investigation of this.






.