[reportlab-users] Large tables - exponentially increasing run times

[John Pywtorak]

It seems to me that you should be able to take any large table and use many 
smaller tables.  Now I realize you discussed how split is implemented now and 
this is what it does.  However, I can say that I produce 7 to 10 page reports 
that are basically one large table.  But, using platypus I just stack table 
after table on top of each other by appending them to the story, or flowable 
array.   And the best part is for me there is no performance issue like you 
mention below.  Platypus seems very adept at adding small tables quickly even 
without row heights.

In the output you would think it is one large table, but it really is not.   I 
had a case where one of the dynamic tables was growing very large, so...

if not row % 4:
    # append current table to story and start a new tables rows

So, I am not really understanding a need for a very large table.  Is this 
approach possible for you?  This is an approach replicated in the HTML world 
especially when you don't want the person to see a blank page for a long time 
in a browser then bam it appears.

Johnny P


On Friday 14 May 2004 10:55 am, Craig Ringer wrote:
> Hi folks
>
> I'm pretty new to ReportLab, but I thought I'd post a few questions
> here, having been unable to find an answer in the list archives or docs.
>
> I've been looking seriously into using ReportLab (specifically,
> platypus) as a print system for an application I'm working on. Part of
> this will involve potentially extremely large table output. My tests
> with "dummy" data in a very simply program are showing some performance
> issues that will be a serious problem for this purpose. I thought I'd
> post my observations so far, then a few questions. If I've missed
> something obvious, then sorry for your time.
*snip*