Suggestions for creating a PDF table

[Kirk Strauser]

Short question:

Is there a good library for generating HTML-style tables with the equivalent
of colspans, automatically sized columns, etc. that can render directly to
PDF?

Longer question:

I'm re-doing a big chunk of locally-written code. I have a
report-generating function that takes a list of lists of lists as input and
returns either a PDF, an HTML table, or an Excel spreadsheet as requested.
For example, input might look like:

makereport('html',
headers=['Invoice number', 'Customer', 'Price'],
data=[
[['123', 'John Doe', '$50.00'],
['Ordered on 2008-01-01 via the website']],
[['124', 'Peter Bilt', '$25.99'],
['Mail via African swallow']]
])

This would result in HTML like:

<table>
<tr>
<th>Invoice number</th>
<th>Customer</th>
<th>Price</th>
</tr>
<tr class="lightbackground">
<td>123</td>
<td>John Doe</td>
<td>$50.00</td>
</tr>
<tr class="lightbackground">
<td colspan="3">Ordered on 2008-01-01 via the website</td>
</tr>
<tr class="darkerbackground">
<td>124</td>
<td>Peter Bilt</td>
<td>$25.99</td>
</tr>
<tr class="darkerbackground">
<td colspan="3">Mail via African swallow</td>
</tr>
</table>

Note particularly how the explanatory notes for each invoice are similar in
appearance to the "primary" report lines they're associated with.

Now, I have a similar transformation to PDF via pdflatex. This works fairly
well but requires a bunch of temp files and subprocesses, and I've never
been 100% happy with the LaTeX output (you have to calculate your own
column widths, for instance). Since I plan to re-write this anyway, I'd
like to find a more widely used library if one was available.

ReportLab seemed *almost* perfect, except that it doesn't support colspans.
As I hope I demonstrated in the example, most of our reports depend on that
ability.

So, again, any thoughts on a PDF generator that can generate tables with the
same kind of flexibility as HTML?

 

[Ken Starks]

Short question:

Is there a good library for generating HTML-style tables with the equivalent
of colspans, automatically sized columns, etc. that can render directly to
PDF?

Longer question:

I'm re-doing a big chunk of locally-written code. I have a
report-generating function that takes a list of lists of lists as input and
returns either a PDF, an HTML table, or an Excel spreadsheet as requested.
For example, input might look like:

makereport('html',
headers=['Invoice number', 'Customer', 'Price'],
data=[
[['123', 'John Doe', '$50.00'],
['Ordered on 2008-01-01 via the website']],
[['124', 'Peter Bilt', '$25.99'],
['Mail via African swallow']]
])


<snip>


Now, I have a similar transformation to PDF via pdflatex. This works fairly
well but requires a bunch of temp files and subprocesses, and I've never
been 100% happy with the LaTeX output (you have to calculate your own
column widths, for instance). Since I plan to re-write this anyway, I'd
like to find a more widely used library if one was available.

<snip>

Short answer: LaTeX should be good. Use XML source; XSLT to TeXML; TeXML
to LaTeX ( uses a python program); compile to PDF.

Longer answer: Can you provide a minimal example of the kind of
LaTeX source you would ideally like ?

If not, your problem is with LaTeX itself, which has if anything
__too_many__ ways of controlling tables rather than inadequate
ways. If so, we may be able to help with the rather arcane
transformation into TeXML format.

As for all the temp files that LaTeX creates, they are easily dealt
with using a makefile or whatever.

Bye for now,
Ken