[OS X TeX] How to fit a table into the letter size paper

Ross Moore ross at ics.mq.edu.au
Tue Dec 30 04:40:16 CET 2008

Hi David, Alain, and others,

On 30/12/2008, at 11:13 AM, David Watson wrote:

> Alain,
> I think most of the of the space savings in this case come from
> a) making the units into math operations and then
> b) using math mode to change the original powers into superscripts,  
> and
> c) positioning the units on a line underneath the heading.

There is also:
  d) the initial  {\footnotesize .... } wrapping around the whole
     tabular, reducing the overall fontsize and baseline skip;
  e) the  \addtolength{\tabcolsep}{-2pt}
     which reduces the amount of space *between* cells in a row.

Of course it is unclear from the original poster's example
what form the data will come in --- presumably numbers,
but with how many digits, and what kind of alignment?
This may affect the widths of columns. In my example code
the header cells dictate the widths, choosing between 3 fixed
amounts which were designed to accommodate the headers.

It may well be that in practice the data dominates in some columns,
which will then affect the overall width of the table.

> The tables in Tufte's "The Visual Display of Quantitative  
> Information" and the documentation for the "booktabs" package  
> provide prime examples of what "tables" should look like.
> As far as defining new math operations, I think I would prefer to  
> use the "SIunits" package in order to facilitate generation of  
> metric units, not to disparage the elegant example that was provided.

Sure; that's certainly a good package to use.

> To the point, the original poster {\em seems} to be asking how to  
> fashion a table that will fit onto letter paper which will be  
> utilized for the recording of data from some sort of physical  
> experiments. In such a case I would prefer a sheet of graph paper,  
> possibly located within a laboratory notebook, or loose leaf in the  
> case of an undergraduate course where the binding is unimportant.

If it was for *recording* data manually, then yes, the small
font-size is not a good idea at all.
I'd assumed that he wants to display data already collected,
or to be collected in future.

With data in a speadsheet, it is easy to add extra columns
in-between, filled-down with &s and \\ at the end of each row.
Then a copy/paste into the LaTeX source can provide the data
cells to populate the {tabular}.

> Given the fact that the data in column 1 (or 0 for you computer  
> science persons) are so miniscule in relation to the column  
> headers, I would opt to either:
> a) use a landscape orientation and format the table as the original  
> poster suggested, or

Certainly using landscape orientation is an option;
  e.g. with a \rotatebox{90}{....} command.
But if you do this, it would be best done within a float,
probably on a page by itself.
It also can help in positioning the \caption to use a
{minipage} environment, estimating the actual width
(which becomes height) required; e.g.,

  }%  end of \rotatebox

Of course there are packages which provide alternative commands
to do some of these things with different markup.
What is important is understanding how the markup relates
to geometrical concepts, and nesting them correctly so that
the coding processes smoothly.

> b) use OpenOffice Calc or MS Excel or some sort of tab-delimited  
> text file for the recording of such data.

Sure. But it doesn't take much to use this for data,
and employ LaTeX for a high-quality, aesthetic layout.

> I could be completely wrong about the intention of the original  
> poster, so understand this would be my approach given what I assume  
> the original assumptions were.
>> Envious regards
>> --schremmer

Hope this helps,


Ross Moore                                       ross at maths.mq.edu.au
Mathematics Department                           office: E7A-419
Macquarie University                             tel: +61 (0)2 9850 8955
Sydney, Australia  2109                          fax: +61 (0)2 9850 8114

More information about the macostex-archives mailing list