finding maximum value in a data set and subtracting from plot

Hi I want to plot a data set, but have it offset (in y) by it's maximum value i.e. in the third plot I have a value at offset. Is there a way of getting gnuplot to look at the data first, determine the maximum value, then subtract it from my data? This would make the difference curve sit below my data, rather than ontop of my data. offset=??? plot pattern u (f($1)):2 t "Observed" with points 1,\ "" u (f($1)):3 t "Calculated" with lines lt rgb "blue",\ "" u (f($1)):(($4)-offset) t "Difference" w lines lt rgb "magenta" I have tried using multiplot with this, but can

Time Hacks or Other Suggestion

I've 6 hours of 1 Hz data that looks like: Time String, X, Y one entry per line in a file. I can plot the X vs Y very nicely using Gnuplot. However, looking at that curve doesn't give any sense of time. One thing I've thought is to try to put a tick mark every 10 minutes. Is there a way to do that with Gnuplot? I'm using version 4.2 What about other suggestions. How do I convey a sense of time for an XY plot?

Color contour plots from "Impossible gnuplot graphs "

Dear all, I am trying to get the color contour script from http://www.phyast.pitt.edu/~zov1/gnuplot/html/contour.html running (see the lower part of this page). While the monochrome script with aligned labels works fine the second script produces errors like if(nr==int(os+w/2) && (d \%2)==0) {a[i]=$1; b[i]=$2; c[i]=$3;} ^ backslash not last character on line Maybe there is only a tiny syntax error or gawk version clash causing this problem. Furthermore, the implementation of the new script and the handling of the new temporary file cont.plt is

Population pyramids

I can't find how to plot a histogram with horizontal boxes. The script below does fine for a population pyramid, printing it on paper only requires to rotate the paper by -90 degrees! It appears as if horizontal boxes are not possible? Thanks in advance, Alex van der Spek +++++++++++++++++ #! usr/bin/env gnuplot #Script to plot population pyramid reset set term wxt 0 size 640,480 title 'PlotPop.plt' set datafile separator ',' set size square set key outside set style fill transparent solid 0.25 set style data histograms set style histogram rowstacked gap 5 set key autotit

sourceforge problem

i'm trying to download gnuplot-4.4.0.tar.gz from sourceforge, but it always gives me the pfd formatted FAQ in place of the required compressed archive

objects with dashed borders

Hello, is it possible to draw an object (like a rectangle or a circle) with a dashed border? When I try set term postscript eps dashed set object circle at 0,0 size 1 fillstyle transparent border lt 2 the border of the circle is a solid line, even though lt 2 should be dashed. I'm using gnuplot 4.5 on Ubuntu. TIA, Michael.

Problems with ytics on multiplot + other questions

Hi I am trying to set up a script to automate images of diffraction data and the Rietveld fits. What I have to plot are the following; Observed points. Calculated points. Difference plot. hkl reflection markers. To do this I have a reasonable script (I think), but still a few issues bugging me. On the difference plot I think there are too many ytics (rather cramped). How do I force gnuplot to use less automatically (the difference plots could change depending on different data sets and quality of refinements)? The hkl reflection markers are manually adjusted to be positione

Offset with yzeroaxis?

I already use set xzeroaxis to define a visible line at y=0. So the x-axis imposes no problem. But the y-axis... My x-data is date information, and I'd like to draw a vertical line at some date x-points. I have not seen the possibility to define "a virtual zero" offset for yzeroaxis at some point in time. Is it correct to assume that I have to issue a separate plot command to achieve that, and there's no offset feature for that? Since auto-scaling shall be performed on the y-range, how can I define the endpoints of the line to match the auto-scaling bounds? Janis

sorting in gnuplot

Dear newsgroup, I need to sort a bunch of numbers in gnuplot. I see this is not possible (at the moment) in gnuplot. Does any gnuplot developer sees a chance for implementing this? No need for a special sorting algorithm. The quicksort algorithm (C.A.R. Hoare), present in the clib will do. Best regards, Francky

Labelling fit results

Would it be possible to read back from fit the standard deviation of the fitted parameters? It is easy to get the parameter values themselves but the sdevs and correlations seem to be available only in the fit.log file. I would like to label my plot with the results. Like this: ############################################################ #! usr/bin/env gnuplot #Example script for plot and fit data reset f(x)=a*x+b a=1 b=1 set datafile separator ',' fit f(x) '$0' using 1:2:(1) via a,b set key outside set size square set grid set label 1 sprintf('Slope = %4.2f +/- %4.2f',a,

WGnuPlot 4.4 on Windows 7 Professional 32-bit

Dear GnuPlot-er, I started using gnuplot in 1991, and from time to time I still use gnuplot. First of all, thanks to all that have kept GnuPlot alive. As I want to use GnuPlot again, I installed the latest binary for the latest release 4.4 on Windows 7 Prof. I am experiencing severe problems: -- The help files are no longer supported (with Windows XP no pro in the last few years) -- I installed the old Help-Viewer (6.1), but still run into problems. The help file cannot be opened. -- The default "wxt" terminal does not show any result. I have to manually kill the pro

Date and time annotations at x-axis, offsets, and bounds

Newbie here in c.g.a.g and only occasional gnuplot user, so please bear with me if I use inappropriate terms. I'm plotting data from a file with a format like 05.04.2010 16:00 0 0 06.04.2010 00:00 11 1 06.04.2010 08:00 18 3 06.04.2010 16:00 2 0 07.04.2010 00:00 10 0 Column 1 and 2 is date/time information (x), column 3 data (y), and column 4 data (y2). So far so good. I'm using gnuplot code like this (sample reduced to x attributes) set xdata time set format x "%a/%d" set timefmt "%d.%m.%Y %H:%M" set xlabel "..." set xtics 86400

Multiple lines from a single data column

I have a data file which has 4 columns, the value to be plotted, an un- needed value, the year and the month-day field: 134 55 2009 02-24 141 55 2010 02-24 146 56 2009 02-25 164 56 2010 02-25 147 57 2009 02-26 97 57 2010 02-26 142 58 2009 02-27 61 58 2010 02-27 I would like to generate a plot which has one separate line for each value in the year column. Whenever I've done multiple line graphs before, it's always been with different columns or data files, and I'm not even sure what I'm looking for to do this, which doesn't help the googling. My dem file so far: set term

Howto update gnuplot from version 4.2 to version 4.4?

I install gnuplot 4.2 patchlevel 5 at my Slackware 12.2. I need update my gnuplot's version, but I don't know how. I install gnuplot by it's own 'make install' process. Help me How do I update gnuplot? I think uninstall gnuplot 4.2 after to install gnuplot 4.4. But at slackware doesn't work removepkg, or slackpkg remove.

Changing margins?

I notice when upgrading from 4.0 to 4.2 that the left margins for the devices I use -- ps, svg, emf, png, x11 and win -- change; they are substantially wider in the newer version. Is this a known issue or something local to me? Pointers please. Thank you, John

Fitting: How to use Poisson error weighting properly

Hi all, in connection with my and a colleague's work with fitting data that have Poisson error bars a dispute has started between us and another colleague in Australia. He raises the issue that the proper weighting should not use the Poisson errors of the data themselves but the Poisson errors of the current Fit trial. This, however, is, as far as I know, not possible in Gnuplot since the errors, if there are any, have to be introduced before the fitting process starts while the "more correct" method would require recalculation of the weights during each iteration. Let y

fit verbosity

Hello, is there a way to decrease the verbosity of the fit command? I´ve got a script that produces a fit.log > 2MB for each run, an i expect it is slowed down a lot by letting all that run over the terminal. The individual fits do usually converge in less that 20 steps, but there are many, many data sets, unfortunately. Most of them are redundant, of course, but it´s hard to know before evaluating them. I expect it won´t be possible without patching!? Another thing: Is there a way to catch a fit error (and restarting the fit with new values afterwards automatically)? Any sug

Multiplot And Overwritten Borders And Tics

Hi! I have written a simple gnuplot script (see code below) to plot files through multiplot mode. It draws two figures, one at the top, and another at the bottom. The data set I am using is this one: http://www.robjhyndman.com/TSDL/data/melbmax.dat The figure at the bottom is just the raw data without any kind of transformations. The figure at the top is the same data file, but transformed through smoothing techniques available in gnuplot: There are a cubic spline (cspline) approximation and a Bezier fitting for the same data set. The problem I have when plotting the Bezie

jgnuplot looking for new maintainer

Hi all, Since my Java days are over, I'm looking for a new maintainer for jgnuplot; a Java library for interfacing with gnuplot. For more information, see: http://jgnuplot.sourceforge.net/ and https://sourceforge.net/projects/jgnuplot/ On average, 30 downloads per week take place. Not a lot but at least a sign that it would be a shame to abandon this project. Send me a message if you are interested in becoming the new maintainer / project manager for jgnuplot. Regards, Pander

514914 Up to date Net and Tech news, knowledge, empowerment 39

I fell in love with the net and all the compuiter teahcnology in the very early years and have loved it ever since, I have an IT Degree and Communications degree and will be working on my doctorate for the next five years, in bwetween I design websites, databases, Multimedia Produxction or anything that wil make me a penny, just student scraping by. But Guess what? I don't want your money, your e-mail, or anything. Each of these sites has a specific topic but they have them style and layout sothey are easy to navigate and ther a search function should take you to exactly where youwant to, o

Unexpected interaction of impulses in 3D with hidden3d

Hi, [Using CVS version 4.5, just compiled from updated source] I have been investigating the 'with impulse' style in 3D. I have data whose z coordinates vary between a positive value and a small negative value. Minimal example file 'heights.dat': 1 2 3 2 3 1 3 2 4 1 3 -1 If I do set xyplane at 0 splot "heights.dat" u 1:2:3 w impulses lw 4 pal then I get the expected (nice) output: each impulse finishes on the xyplane (z=0). However, if I now enable hidden3d: set hidden3d then the impulses are correctly hidden, but extend too far, apparently down to the minimum

Plotting data from a file

Hi, I'm using gnuplot 4.4.0 on OS X 10.6.2 and am trying to plot the file ecut.txt which looks like #Ecut Length Atomisation Energy 10 0.805312626 0.165625435 15 0.794875914 0.171276982 20 0.782939689 0.173722651 25 0.775818652 0.174689776 30 0.772 0.175219486 35 0.772614854 0.175525673 40 0.772084357 0.175793694 45 0.771989336 0.176002384 50 0.771522545 0.176162818 with the command (e.g) plot 'ecut.txt' using 1:2 and get the response warning: Skipping unreadable file "ecut.txt" No data in plot What am I doing wrong? Cheers.

Log scale and image data

Hello! I have a data set which I am trying to plot with the latest CVS version of GNUplot that will not plot with log scaling. The error message I get with autoscaling is ""/tmp/data.gnu", line 49174: y range must be greater than 0 for log scale" and it is! If I hard code the range to something like : set yrange [10:200], I don't get the error message, but I don't get a plot either. I am using a very simple plot line: plot "-" using 1:3:5 axes x1y1 with image and the data follows it. If I go to linear scale, I get my plot as well. Is this a bug? I'm happy to send a data fil

new version 4.4 slow to start on vista

hi, i installed the new version today and found it very slow to start on windows vista 32bit. i have microsoft security essentials installed. the actual exe opens right away but once i select one of my plt files to open it takes about a minute to display the data. after this one time slow down the program is fast.

How to change the scale of the colorbox?

Hi, everyone! The problem about my plotting is that the scale of the colorbox on the right seems ugly. I just want to adjust the scale as 0, 10, 50, 100, 1000, 4000, 8000. How can I set the related parameters for this plot? Thanks in advance! Below is my Gnuplot script. set xrange [-8:8] set yrange [-8:8] set xlabel "X" set ylabel "Y" set isosample 80 set contour base set cntrparam levels discrete 1.0, 5.0, 10.0, 20.0, 30.0, 50.0, 80.0, \ 120.0, 200.0, 330.0, 500.0, 700.0, 920.0, 1200.0, 1500.0, 2000.0, 2500.0, \ 3200.0, 4000.0, 5000.0, 8000.0 unset clabel unset key un