Is is possible to incrementally write an Excel document created u sing POI/HSSF?

[Dana Spiegel <da...@farechase.com>]

I'm programmatically generating an Excel document that contains a lot of
data (the final document size is 20+ MB). I would like to incrementally
flush the document to disk to conserve memory usage (so the entire document
doesn't have to kept in memory through the POI package).

Is this possible? Is the POI/HSSF package efficient in its memory usage such
that such a thing isn't necessary?

Any insight would be helpful; I've searched the archives and can't seem to
find any discussion of this issue...

Thanks,

D a n a   S p i e g e l 
FareChase  ::  www.farechase.com 
50 West 23rd Street, PH, New York, NY 10010 
p  +1 646 452 0665  ::  m  +1 917 402 0422  ::  f +1 646 452 0667  ::  e
dana@farechase.com

Re: Is is possible to incrementally write an Excel document created u sing POI/HSSF?

["Andrew C. Oliver" <ac...@apache.org>]

However anyone who wants to help, can! Grab the performance-branch tag!

cvs co -r performance-branch jakarta-poi

-Andy

Avik Sengupta wrote:
> If Only!
> 
> But no, we cannot incrementally flush a document to disk, since the file
> format contains upstream pointers to downstream data! So we have to hold
> the whole file in memory while writing.
> 
> POI does use a significant amount of memory to do its work. Some of it,
> as i said above, is a function of how the file format works. There is
> some effort going on to refactor the code to use less memory, but thats
> in its very early stages. 
> 
> -
> avik
> 
> 
> On Mon, 2003-01-13 at 23:19, Dana Spiegel wrote:
> 
>>I'm programmatically generating an Excel document that contains a lot of
>>data (the final document size is 20+ MB). I would like to incrementally
>>flush the document to disk to conserve memory usage (so the entire document
>>doesn't have to kept in memory through the POI package).
>>
>>Is this possible? Is the POI/HSSF package efficient in its memory usage such
>>that such a thing isn't necessary?
>>
>>Any insight would be helpful; I've searched the archives and can't seem to
>>find any discussion of this issue...
>>
>>Thanks,
>>
>>D a n a   S p i e g e l 
>>FareChase  ::  www.farechase.com 
>>50 West 23rd Street, PH, New York, NY 10010 
>>p  +1 646 452 0665  ::  m  +1 917 402 0422  ::  f +1 646 452 0667  ::  e
>>dana@farechase.com
>>
> 
> 
> 
> 
> 
> --
> To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
> For additional commands, e-mail: <ma...@jakarta.apache.org>
> 
> 

Re: Is is possible to incrementally write an Excel document created u sing POI/HSSF?

[Avik Sengupta <av...@apache.org>]

If Only!

But no, we cannot incrementally flush a document to disk, since the file
format contains upstream pointers to downstream data! So we have to hold
the whole file in memory while writing.

POI does use a significant amount of memory to do its work. Some of it,
as i said above, is a function of how the file format works. There is
some effort going on to refactor the code to use less memory, but thats
in its very early stages. 

-
avik


On Mon, 2003-01-13 at 23:19, Dana Spiegel wrote:
> I'm programmatically generating an Excel document that contains a lot of
> data (the final document size is 20+ MB). I would like to incrementally
> flush the document to disk to conserve memory usage (so the entire document
> doesn't have to kept in memory through the POI package).
> 
> Is this possible? Is the POI/HSSF package efficient in its memory usage such
> that such a thing isn't necessary?
> 
> Any insight would be helpful; I've searched the archives and can't seem to
> find any discussion of this issue...
> 
> Thanks,
> 
> D a n a   S p i e g e l 
> FareChase  ::  www.farechase.com 
> 50 West 23rd Street, PH, New York, NY 10010 
> p  +1 646 452 0665  ::  m  +1 917 402 0422  ::  f +1 646 452 0667  ::  e
> dana@farechase.com
>