|
From: sfeam <eam...@gm...> - 2014-11-09 18:39:44
|
On Monday, 10 November 2014 05:15:38 AM Dave Horsfall wrote: > I know, it's mathematically impossible (it tends to disappear towards -ve > infinity), but bear with me. > > Behold the graph over on www.horsfall.org/spamlog.pdf (about 100k); it's a > daily plot of my email logs, and amongst other things it includes the spam > that was actually delivered (shown in pink, of course, as opposed to the > spam that was rejected in red). I had to use a logarithmic scale because > of the magnitudes involved, which means around 3rd November was a > spam-free day (yes, I was surprised as well!). > > It looks ugly, so I would like to re-map "0" into "0.1", but I cannot > think of an elegant way of doing this. An explicit axis label overrides an auto-generated one, so this should do the trick: set ytics add ("0.0" 0.1) plot ... Ethan > > 1) Conditional expressions in the "plot" command; yuck. > > 2) Something like "if ($7 == 0) $7 = 0.1" before the plot; didn't work. > > 3) Playing with "set zero"; I must've misunderstood how it's used. > > 4) A pre-filter on the data; that amounts to tampering with it. > > So, is there an elegant way to redefine "0" as "0.1", or do I have to > pollute the "plot" command? > > Thanks. > > |