<br><br><div><span class="gmail_quote">2007/2/23, Thomas Friedrichsmeier <<a href="mailto:thomas.friedrichsmeier@ruhr-uni-bochum.de">thomas.friedrichsmeier@ruhr-uni-bochum.de</a>>:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
On Friday 23 February 2007 21:47, Thomas Friedrichsmeier wrote:<br>> If there are, then it may still be worth considering,<br>> whether the type of data I suggested should not be supported in some way as<br>> well, perhaps with a "tabulate" option.
<br><br>Pseudo-code to make my idea a bit more clear:<br><? if (getRK_val ("tabulate)) { ?><br>rk.temp.x <- table (<? echo ($var); ?>, exclude=NULL)<br><? } else { ?><br>rk.temp.x <- <? echo ($var); ?>
<br>if (!is.numeric (rk.temp.x)) {<br> # perhaps use rk.print() instead of warning()<br> warning ("Data may not be numeric, but proceeding as requested.\nDid you<br>forget to check the tabulate option?")
<br>}<br><? } ?><br><br>I think an advantage is, that there is less "magic" going on outside the<br>user's control. Currently factors are simply dealt with differently than<br>vectors, and I think this is what has been disturbing me in the first place.
<br>The user will not have any way to see why one object gets tabulated, but the<br>other doesn't, without looking at the generated code (and understanding the<br>differentiation between factor and vector in R, which is different from some
<br>other statistics tools). This is not a problem, as long as the "magic" always<br>does "the right thing" - but I'm afraid it will not always read the user's<br>mind correctly, so let's rather use a good old checkbox, instead.
<br></blockquote></div><br>Yes it seems good. I will try it. Probably this apply to the barplot too or should we leave it as is?<br>