I wonder if there is a better way to manage xlabels in particular in a case when a plot has several datasets with non-matching Xs.
xlabels apparently correspond to the first dataset of a plot. This means that there can be too many of them to be readable and also the range of the first dataset may not be wide enough to represent Xs for the rest of the data. My best shot (see below) is to provide an additional dataset representing labels and their positions first, but the dummy dataset goes into the plot legend which is a bit of nuisance when the plot is to be shared. Any hints concerning this? load'plot' pd 'reset' pd 'key "xlabel dataset","2*1 3 5" "3*2 3 6 7"' NB. dummy dataset will turn up in the legend pd 'color black,green,red' NB. set colors explicitly, assign black to xlabel ds so it will stick out less pd 'xlabel "one" "three" "five" "seven"' NB. the labels xl =. 1 3 5 7 NB. Xs for the xlabel set pd xl;0#~#xl NB. first submit xlabel dataset -- need to provide Ys too, let them be zero x1=. 1 3 5 NB. the actual data x2=. 2 3 6 7 pd x1;2*x1 pd x2;3*x2 pd 'show' regards, Danil ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
