> Friedrich Huebler wrote:
>
> You can evaluate the range of the variables in your graph within the
> do-file to identify appropriate labels.
>
> . sysuse auto
> . sum weight, meanonly
> . local xmin = floor(r(min) / 500) * 500
> . local xmax = ceil(r(max) / 500) * 500
> . sum mpg, meanonly
> . local ymin = floor(r(min) / 5) * 5
> . local ymax = ceil(r(max) / 5) * 5
> . scatter mpg weight, xscale(r(`xmin' `xmax'))
> xlabel(`xmin'(500)`xmax') yscale(r(`ymin' `ymax'))
> ylabel(`ymin'(5)`ymax')
>
> I selected the values 5 and 500 after visual inspection of the data.
> You can also write code that specifies appropriate steps for the axis
> labels based on the difference between the maximum and minimum values
> of a variable. The values in the example above would then be replaced
> by two additional macros, e.g. `xstep' instead of 500 and `ystep'
> instead of 5.
Yes, it's good to know that it is doable. But my point is that the
user should not need to do all this extra work just to turn off the
extra margin, without having to know the (usually reasonably good)
default label positions in the interior of the axis.
(Why would one want to turn off the extra margins? In my real
application, the x-axis is a %tc variable and it adds way too much
margin, presumably because the %tc variable measured in milliseconds
are huge numbers, extending the axis into the next day.)
h.
--
+---
| Hiroyuki Kawakatsu
| Business School, Dublin City University
| Dublin 9, Ireland. Tel +353 (0)1 700 7496
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
