It would increase their memory consumption but would have a next to null
effect on performance (just the time difference between dimensioning a
larger array and smaller array).

However I would seriously urge you to prevent generation of this size. 

-Andy

On Thu, 17 Oct 2002 11:10:54 +0000, Clifton.Crai wrote:

> 
> Just one small question. Would adjusting the constants at the top of Sheet
> and Workbook have any adverse affects on smaller spreadsheets? In other
> words would this cause smaller spreadsheets to execute longer? I'm just
> wondering if it would be worth it to bump these up a little in anticipation
> of the dumb user with the big extract files.
> 
> Cliff
> 
> 
> 
> "Andrew C. Oliver" <[EMAIL PROTECTED]>
> 10/17/2002 10:34 AM
> 
> 
> 
> 
> 
> Please respond to "POI Users List" <[EMAIL PROTECTED]>
> 
> To:    POI Users List <[EMAIL PROTECTED]>
> cc:
> 
> Subject:    RE: POI Speed...
> 
> 
> My technical opinions (take them for what you will) on this:
> 
> 1. HSSF 2.0 should not be used to generate spreadsheets of >6mb in
> length.  (mid 3.0 I'll probably say 8mb)
> 
> 2. Many versions of Excel have great difficulty loading spreadsheets of
>>4mb in length.  (Once we write INDEX/DBCELL records, mid-3.0 I'll
> probably say 6-8mb)
> 
> 3. This file format is NOT an efficient transfer format and should not
> be used as such.  There are many technical reasons for this.
> 
> HSSF 2.0's sweetspot is a sheet of 150k-1mb.
> 
> You can make performance adjustments by increasing the size of the seed
> constants at the top of Sheet and Workbook, etc and recompiling.  This
> will seed ArrayLists and other collections with larger numbers (so that
> garbage collections happens less, and array copies happen less).
> 
> -Andy
> 
> On Thu, 2002-10-17 at 09:49, [EMAIL PROTECTED] wrote:
>>
>> I'm not sure what part is the biggest bottleneck. I just returned from a
>> United Way meeting and noticed it was finished. The spreadsheet turned
> out
>> to be 14.1MB and it's taking roughly a couple of minutes to download from
>> our AS/400. (Our AS/400 is actually hosted some 30+ miles away from our
>> office and our network connection isn't the best.) I actually miscounted
>> the columns (I didn't window over to the right). It was really a 90
> column
>> table with 11297 rows so I'm looking at writing 1,016,730 cells of data
>> plust the extra 90 column headers which brings this total to 1,016,820. A
>> pretty sizable file for a spread sheet yet I still don't think I should
>> have gotten the results that I got. It took over 5 minutes to create.
> I'll
>> try to bench mark the building of the result set and compare this with
> the
>> time it takes to loop through all records and write to a spreadsheet.
>>
>> Cliff
>>
>>
>>
>> "Iouri Michtchenko" <[EMAIL PROTECTED]>
>> 10/17/2002 09:00 AM
>>
>>
>>
>>
>>
>> Please respond to "POI Users List" <[EMAIL PROTECTED]>
>>
>> To:    "POI Users List" <[EMAIL PROTECTED]>
>> cc:
>>
>> Subject:    RE: POI Speed...
>>
>>
>> Are you sure that the POI part is the reason?
>> Why not to put some debug output???
>> I use POI for almost the same kind of things and I don't have any
> problems
>> with performance.
>>
>> -----Original Message-----
>> From: [EMAIL PROTECTED] [mailto:Clifton.Craig@;rrd.com]
>> Sent: Thursday, October 17, 2002 3:03 PM
>> To: POI Users List
>> Subject: Re: POI Speed...
>>
>>
>>
>> It's still running! Also note that my program is basically creating a
> JDBC
>> result set from a SQL statement and writing this data column by column
> row
>> by row to the spreadsheet. Nothing fancy. It's still running!
>>
>> Cliff
>>
>>
>>
>> --
>> To unsubscribe, e-mail:   <mailto:poi-user-unsubscribe@;jakarta.apache.org
>>
>> For additional commands, e-mail: <mailto:poi-user-help@;jakarta.apache.org
>>
>>
>>
>>
>> --
>> To unsubscribe, e-mail:   <mailto:poi-user-unsubscribe@;jakarta.apache.org
>>
>> For additional commands, e-mail: <mailto:poi-user-help@;jakarta.apache.org
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> --
>> To unsubscribe, e-mail:   <mailto:poi-user-unsubscribe@;jakarta.apache.org
>>
>> For additional commands, e-mail: <mailto:poi-user-help@;jakarta.apache.org
>>
>>
> --
> http://www.superlinksoftware.com - software solutions for business
> http://jakarta.apache.org/poi - Excel/Word/OLE 2 Compound Document in
> Java
> http://krysalis.sourceforge.net/centipede - the best build/project
> structure
>       a guy/gal could have! - Make Ant simple on complex Projects!
> The avalanche has already started. It is too late for the pebbles to
> vote.
> -Ambassador Kosh
> 
> 
> --
> To unsubscribe, e-mail:   <mailto:poi-user-unsubscribe@;jakarta.apache.org>
> For additional commands, e-mail: <mailto:poi-user-help@;jakarta.apache.org>




--
To unsubscribe, e-mail:   <mailto:poi-user-unsubscribe@;jakarta.apache.org>
For additional commands, e-mail: <mailto:poi-user-help@;jakarta.apache.org>

Reply via email to