从同一文件中绘制多个系列 [英] Plot multiple series from the same file

问题描述

我有一个类似于以下内容的数据文件:

I've got a data file that looks something like the following:

2016-08-16 dogs 12
2016-08-16 cats 9
2016-08-17 cats 4
2016-08-17 mice 42
2016-08-17 dogs 6

即:date series count.

请注意，该系列并非每天都有一致的顺序(并且可能根本不会出现).

Note that the series are not in a consistent order on each day (and might not appear at all).

此外，我的输入文件也没有按顺序排列日期(尽管如果我必须可以用sort进行修复)，并且某些日期可能会丢失(因为当天没有数据点.

Further, my input file doesn't have the dates in order either (though I can fix that with sort if I have to), and some of the dates might be missing (because there are no data points for that day).

我想使用 gnuplot 进行渲染. X轴应显示日期； Y轴应显示计数；每个系列应该是自己的一行，并在图例中带有标签.

I'd like to use gnuplot to render this. The X-axis should display the date; the Y-axis should display the count; each series should be its own line, with labels in the legend.

我知道如何从一个单独的文件中渲染每个系列，但是该系列不是固定的，因此我正在寻找不需要该系列的东西.

I know how to render each series from a separate file, but the series are not fixed, so I'm looking for something that doesn't require that.

我该怎么做?

注意:我正在使用gnuplot 4.4

推荐答案

也许首先可以这样转换输入文件(input.txt)文件:

perhaps one could first transform the input file (input.txt) file like so:

sort -k2,2 -k1,1 input.txt | \
gawk '{if($2!=prev){if(NR>1){print "\n";}print "date", $2; prev=$2;}print $1,$3}' > input2.txt

在这种情况下提供的

which in this particular case provides

date cats
2016-08-16 9
2016-08-17 4


date dogs
2016-08-16 12
2016-08-17 6


date mice
2016-08-17 42

，然后使用

datafile="input2.txt"
stats datafile

set xdata time
set timefmt '%Y-%m-%d'

N=int(STATS_blocks)-1
plot for [i=0:N] datafile index i using 1:2 w lp t columnhead(2)

产生

这篇关于从同一文件中绘制多个系列的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！