gnuplot-info — For questions and discussion about gnuplot

Re: [Gnuplot-info] Conditional ignoring lines in data file

[Thomas S. <t.s...@fz...>]

it means 'undefined value' and it's gnuplot specific.
see:
http://www.gnuplot.info/docs/gnuplot.html#x1-2900013.2.3
http://www.gnuplot.info/docs/gnuplot.html#x1-15100065.2.9


fusselpelusa wrote:
> 
> Perfect, that works very well.
> THX!
> I am wandering about the syntax 0/0. Obviously it means don't plot 
> anything. is this a C command or a gnuplot internal code?
> 
> Thomas
> 
> Thomas Sefzick wrote:
>> plot 'test.dat' using 3:($2<0?$2:0/0) w l
>>
>>
>> fusselpelusa wrote:
>>   
>>> Hello,
>>> is there a way to ignore lines in a data file if a condition is 
>>> fulfilled, and still have the rest connected by a line? Ideally the gap 
>>> between the ignored data points would not be connected with a line.
>>>
>>> I understand that I can prepare my data file differently, but I will 
>>> then have a connection over the gaps and I thought there could be also a 
>>> direct way in gnuplot. Maybe with a cat or awk command inside the plot 
>>> command? I am unfortunately not too familiar with their syntax.
>>> I have quite a few gaps in my data file and plotting each line as a 
>>> multiplot in one diagram is not practical.
>>>
>>> I was thinking of something like this, but the syntax is of course not 
>>> right:
>>>
>>> plot 'test.dat' ($2)<0 ? (u 3:2 w l) :
>>>
>>> Thanks
>>> Thomas
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Download Intel&#174; Parallel Studio Eval
>>> Try the new software tools for yourself. Speed compiling, find bugs
>>> proactively, and fine-tune applications for parallel performance.
>>> See why Intel Parallel Studio got high marks during beta.
>>> http://p.sf.net/sfu/intel-sw-dev
>>> _______________________________________________
>>> gnuplot-info mailing list
>>> gnu...@li...
>>> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
>>>
>>>
>>>     
>>
>>   
> 
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 
> 

-- 
View this message in context: http://old.nabble.com/Conditional-ignoring-lines-in-data-file-tp27929592p27942771.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Conditional ignoring lines in data file

[Thomas P. <jun...@gm...>]

Perfect, that works very well.
THX!
I am wandering about the syntax 0/0. Obviously it means don't plot 
anything. is this a C command or a gnuplot internal code?

Thomas

Thomas Sefzick wrote:
> plot 'test.dat' using 3:($2<0?$2:0/0) w l
>
>
> fusselpelusa wrote:
>   
>> Hello,
>> is there a way to ignore lines in a data file if a condition is 
>> fulfilled, and still have the rest connected by a line? Ideally the gap 
>> between the ignored data points would not be connected with a line.
>>
>> I understand that I can prepare my data file differently, but I will 
>> then have a connection over the gaps and I thought there could be also a 
>> direct way in gnuplot. Maybe with a cat or awk command inside the plot 
>> command? I am unfortunately not too familiar with their syntax.
>> I have quite a few gaps in my data file and plotting each line as a 
>> multiplot in one diagram is not practical.
>>
>> I was thinking of something like this, but the syntax is of course not 
>> right:
>>
>> plot 'test.dat' ($2)<0 ? (u 3:2 w l) :
>>
>> Thanks
>> Thomas
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Download Intel&#174; Parallel Studio Eval
>> Try the new software tools for yourself. Speed compiling, find bugs
>> proactively, and fine-tune applications for parallel performance.
>> See why Intel Parallel Studio got high marks during beta.
>> http://p.sf.net/sfu/intel-sw-dev
>> _______________________________________________
>> gnuplot-info mailing list
>> gnu...@li...
>> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
>>
>>
>>     
>
>   

Re: [Gnuplot-info] Gnuplot - plot all files in folder with a certain signature (using the asterisk)

[Thomas S. <t.s...@fz...>]

plot for [fn in system("ls *cat*.dat")] fn using 1:2 title fn

if not using a unix-like operating system replace 'ls' with
the 'listing filenames' command of your operating system.


mean6 wrote:
> 
> Hello everybody,
> 
> I am new to this forum - as well as to gnuplot.
> Please allow me a question:
> 
> Is it possible to tell gnuplot by a simple command to plot in one graph
> all files that contain a certain text in their filename?
> 
> Something like that:
> 
> plot *cat*.txt
> 
> which would plot all files in the actual folder that contain the word
> "cat" in their filename.
> 
> Thanks for your help folks!
> 
> 

-- 
View this message in context: http://old.nabble.com/Gnuplot---plot-all-files-in-folder-with-a-certain-signature-%28using-the-asterisk%29-tp27937745p27937808.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

[Gnuplot-info] curve on top of a contour plot

[Timothy C. <tim...@ma...>]

Hi

I have made a shaded contour plot from data of the for F(x,y) and i would like to plot a 1d curve  y=g(x) on top of this. When i try just using multiplot the axies are miss aligned and the plots are different sizes.
Is there any way to do this?

Cheers

Re: [Gnuplot-info] Gnuplot 4.2 - problem manipulating "Seconds since Unix Epoch" timestamps

[Thomas S. <t.s...@fz...>]

gnuplots convert the 'seconds since unix epoch' into
'seconds since gnuplot epoch' which started on 01/01/2000,
that's why you see x-values like 3.2 * 10^8

so the values you specify for x-range should be in 'seconds
since 01/01/2000,00:00:00'


Pieter Thysebaert-3 wrote:
> 
> Dear gnuplot experts,
> 
> I'm running Debian Lenny with gnuplot version
> 
> G N U P L O T
> Version 4.2 patchlevel 2
> last modified 31 Aug 2007
> System: Linux 2.6.26-1-amd64
> 
> My problem: I want to chart "sar" output, using "sar -h -P ALL -f
> datafile.sar | grep "%idle" > data.csv"
> 
> I'm using the "-h" option to get sar to record time using "seconds since
> Unix Epoch" format. There's only one CPU in the machine, so each line has
> a
> unique timestamp in reality.
> 
>>From gnuplot, I'm using the following formatting
> 
> set xdata time
> set timefmt "%s"
> set format x "%D\n%R"
> 
> # plot CPU usage, timestamp is column 3, values are column 6
> plot 'data.csv' using 3:(100.0-$6) w lp
> 
> I'm using a sample set spanning from March 10 to March 17 2010.
> 
> The epoch seconds are seemingly correctly interpreted by gnuplot, their
> values are in the 1.2 * 10^9 range and I've verified them using i.e.
> 
> date --date="Jan 1, 1970 00:00:00 +0000 +1268694000 seconds"
> 
> which returns
> 
> Tue Mar 16 00:00:00 CET 2010
> 
> Now, on to the problem: I want to limit the chart to the xrange
> corresponding to, for instance, Tue 16 March 2010
> 
> Observation 1: when hovering over the "wxt" terminal graph, shown x-values
> are in the range 3.2 * 10^8 (way smaller than the actual x-values in the
> file)
> 
> Observation 2:
> 
> set xrange [1268694000:12686940060]
> replot
> 
> I expect the above (using the example dates from above) to plot one minute
> of data from Tue Mar 16 2010.
> 
> However, the last "replot" command returns an "all points y value
> undefined
> !" error.
> 
> When manually setting the yrange to [0:100], I get another "clue":
> 
> set yrange [0:100]
> replot
> 
> shows dates on my x-axis in the year 2040; more precisely, it shows 23:00
> to
> 23:01 on March 14, 2040
> 
> Using a "date" conversion, those dates should correspond to a number of
> 2.2*10^9 seconds since Unix epoch.
> 
> 
> I'm completely clueless at the moment: can someone point out my error, or
> tell me how I can correctly limit x-values when "epoch seconds" is used as
> X
> time format?
> 
> 
> Thanks,
> Pieter
> ------------------------------------------------------------------------------
> Download Intel® Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 
> 

-- 
View this message in context: http://old.nabble.com/Gnuplot-4.2---problem-manipulating-%22Seconds-since-Unix-Epoch%22-timestamps-tp27930349p27932452.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Conditional ignoring lines in data file

[Thomas S. <t.s...@fz...>]

plot 'test.dat' using 3:($2<0?$2:0/0) w l


fusselpelusa wrote:
> 
> Hello,
> is there a way to ignore lines in a data file if a condition is 
> fulfilled, and still have the rest connected by a line? Ideally the gap 
> between the ignored data points would not be connected with a line.
> 
> I understand that I can prepare my data file differently, but I will 
> then have a connection over the gaps and I thought there could be also a 
> direct way in gnuplot. Maybe with a cat or awk command inside the plot 
> command? I am unfortunately not too familiar with their syntax.
> I have quite a few gaps in my data file and plotting each line as a 
> multiplot in one diagram is not practical.
> 
> I was thinking of something like this, but the syntax is of course not 
> right:
> 
> plot 'test.dat' ($2)<0 ? (u 3:2 w l) :
> 
> Thanks
> Thomas
> 
> 
> 
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 
> 

-- 
View this message in context: http://old.nabble.com/Conditional-ignoring-lines-in-data-file-tp27929592p27932059.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

[Gnuplot-info] Gnuplot 4.2 - problem manipulating "Seconds since Unix Epoch" timestamps

[Pieter T. <pie...@gm...>]

Dear gnuplot experts,

I'm running Debian Lenny with gnuplot version

G N U P L O T
Version 4.2 patchlevel 2
last modified 31 Aug 2007
System: Linux 2.6.26-1-amd64

My problem: I want to chart "sar" output, using "sar -h -P ALL -f
datafile.sar | grep "%idle" > data.csv"

I'm using the "-h" option to get sar to record time using "seconds since
Unix Epoch" format. There's only one CPU in the machine, so each line has a
unique timestamp in reality.

>From gnuplot, I'm using the following formatting

set xdata time
set timefmt "%s"
set format x "%D\n%R"

# plot CPU usage, timestamp is column 3, values are column 6
plot 'data.csv' using 3:(100.0-$6) w lp

I'm using a sample set spanning from March 10 to March 17 2010.

The epoch seconds are seemingly correctly interpreted by gnuplot, their
values are in the 1.2 * 10^9 range and I've verified them using i.e.

date --date="Jan 1, 1970 00:00:00 +0000 +1268694000 seconds"

which returns

Tue Mar 16 00:00:00 CET 2010

Now, on to the problem: I want to limit the chart to the xrange
corresponding to, for instance, Tue 16 March 2010

Observation 1: when hovering over the "wxt" terminal graph, shown x-values
are in the range 3.2 * 10^8 (way smaller than the actual x-values in the
file)

Observation 2:

set xrange [1268694000:12686940060]
replot

I expect the above (using the example dates from above) to plot one minute
of data from Tue Mar 16 2010.

However, the last "replot" command returns an "all points y value undefined
!" error.

When manually setting the yrange to [0:100], I get another "clue":

set yrange [0:100]
replot

shows dates on my x-axis in the year 2040; more precisely, it shows 23:00 to
23:01 on March 14, 2040

Using a "date" conversion, those dates should correspond to a number of
2.2*10^9 seconds since Unix epoch.


I'm completely clueless at the moment: can someone point out my error, or
tell me how I can correctly limit x-values when "epoch seconds" is used as X
time format?


Thanks,
Pieter

[Gnuplot-info] Conditional ignoring lines in data file

[Thomas P. <jun...@gm...>]

Hello,
is there a way to ignore lines in a data file if a condition is 
fulfilled, and still have the rest connected by a line? Ideally the gap 
between the ignored data points would not be connected with a line.

I understand that I can prepare my data file differently, but I will 
then have a connection over the gaps and I thought there could be also a 
direct way in gnuplot. Maybe with a cat or awk command inside the plot 
command? I am unfortunately not too familiar with their syntax.
I have quite a few gaps in my data file and plotting each line as a 
multiplot in one diagram is not practical.

I was thinking of something like this, but the syntax is of course not 
right:

plot 'test.dat' ($2)<0 ? (u 3:2 w l) :

Thanks
Thomas

Re: [Gnuplot-info] plotting n:n matrix from .csv with gnuplot

[Thomas S. <t.s...@fz...>]

delete the first line and the first column in your datafile.
then plot with

splot 'datafilename' using (50.+$1*350.):(667.5+$2*2.5):3 matrix with
linespoints


Sinatra wrote:
> 
> Hello,
> 
> I am new to gnnuplot and facing the following problem: I would like to
> plot a 3D surface from a .csv-file, which contains an n:n matrix of data.
> I assume to use the splot command. However, as I am not familiar with
> gnuplot I could not yet figure out if this problem really can be solved
> with gnuplot. 
> 
> You can find an abstract of the data below. 
> 
> 	50.00	400.00	750.00	1100.00	1450.00
> 667.50	4.728	4.647	4.569	4.503	4.438
> 670.00	4.725	4.649	4.567	4.503	4.435
> 672.50	4.724	4.649	4.568	4.501	4.435
> 675.00	4.722	4.649	4.568	4.502	4.433
> 677.50	4.722	4.649	4.569	4.500	4.433
> 680.00	4.720	4.647	4.566	4.496	4.430
> 682.50	4.718	4.647	4.565	4.496	4.429
> 685.00	4.716	4.646	4.562	4.494	4.426
> 687.50	4.716	4.645	4.561	4.492	4.422
> 690.00	4.715	4.645	4.558	4.492	4.422
> 692.50	4.713	4.644	4.556	4.490	4.420
> 
> 
> Is there someone who can give me a hint how to proceed with gnuplot? I
> would really appreciate it!
> 
> Thanks in Advance,
> 
> Sinatra
> 

-- 
View this message in context: http://old.nabble.com/plotting-n%3An-matrix-from-.csv-with-gnuplot-tp27914440p27916524.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] tic mark color

[Thomas S. <t.s...@fz...>]

have a look at the following script:

reset
# border and axis tics red
set border linetype 1
# tic labels in a different color
set xtics textcolor rgb "blue"
set ytics textcolor rgb "green"
# black rectangle drawn on top of red border (works for version >= 4.2.5)
set obj 1 rect from graph 0,0 to graph 1,1 front fillstyle empty
set yrange [-1.1:1.1]
plot sin(x) linetype 3


aaron brick-3 wrote:
> 
> hi all,
> are tic marks always black (in 4.2)? i am making a graph for marketing
> purposes and haven't found a way to control their color other than
> using grid lines, which means no "out" tics nor length control. any
> suggestions will be appreciated!
> 
> thanks,
> aaron.
> 
> 
> --   aaron brick
> --   aa...@li...
> 
> ------------------------------------------------------------------------------
> Download Intel® Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 
> 

-- 
View this message in context: http://old.nabble.com/tic-mark-color-tp27911768p27915244.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

[Gnuplot-info] tic mark color

[aaron b. <bri...@gm...>]

hi all,
are tic marks always black (in 4.2)? i am making a graph for marketing
purposes and haven't found a way to control their color other than
using grid lines, which means no "out" tics nor length control. any
suggestions will be appreciated!

thanks,
aaron.


--   aaron brick
--   aa...@li...

Re: [Gnuplot-info] Gnuplot is including Feb29 even though it is not a leap year

[Thomas S. <t.s...@fz...>]

hardcoding, made easy:

--- src/time.c.orig     2009-12-25 01:30:33.000000000 +0100
+++ src/time.c  2010-03-15 17:02:23.959058370 +0100
@@ -169,3 +169,3 @@
     /* make relative times work (user-defined tic step) */
-    tm->tm_year = ZERO_YEAR;
+    tm->tm_year = ZERO_YEAR+10;
 
but again: mark this modified version to be able to distinguish
it from the unmodified one.


Thomas Sefzick wrote:
> 
> the best way would be to change your parsing script.
> 
> adding 10 years inside the 'using' option of the 'plot' command
> (like 'using (timecolumn(1)+(2*366+8*365)*24*60*60):2')
> doesn't work because the leap year calculation is done before
> adding the 10 years.
> 
> hardcoding could be done, although i would discourage from
> doing this:
> in 'time.c' change ZERO_YEAR and adjust JAN_FIRST_WDAY
> and SEC_OFFS_SYS accordingly, change 'PROGRAM' in 'setshow.h'
> and rename the executable to distinguish it from the official
> version.
> 
> 
> lozwell wrote:
>> 
>> Thanks for the reply. Unfortunately I cannot easily set the year in the
>> data file, as I am just parsing log files. I could however change my
>> parsing script to add a year. Or, is there any way to hard code the year
>> in gnuplot? I tried your way, which produced the same results as 2000 is
>> a leap year.
>> 
>> 
>> 
>> Thomas Sefzick wrote:
>>> 
>>> change x-axis format to
>>> set format x "%b\n%d\n%Y"
>>> and you will see that gnuplot assumes its own 'year zero' (=2000) when
>>> no information about the year is given.
>>> gnuplot can not guess of which year your data are.
>>> you should include the year in your data.
>>> 
>>> 
>>> lozwell wrote:
>>>> 
>>>> Hi, I am having a problem graphing a lot of data using the date/time
>>>> over a 2 week period. It looks as though gnuplot is adding in Feb29 to
>>>> my x axis with no data points, even though there is no Feb29 as 2010 is
>>>> not a leap year. My servers date is the correct year of 2010. This is
>>>> causing a gap in my graph to show a large gap. Any ideas how to fix
>>>> this? As an example, the following code will also produce the same
>>>> results.
>>>> 
>>>> [test.conf]
>>>> set terminal png
>>>> set xdata time
>>>> set xtic 86400
>>>> set timefmt "%b%d-%H:%M:%S"
>>>> set format x "%b\n%d"
>>>> set yrange [0:10]
>>>> set output "test.png"
>>>> set grid
>>>> plot "test.dat" using 1:2 with linespoints
>>>> [/test.conf]
>>>> 
>>>> [test.dat]
>>>> Feb26-01:00:00 1
>>>> Feb26-12:00:00 5
>>>> Feb26-23:00:00 1
>>>> Feb27-01:00:00 1
>>>> Feb27-12:00:00 5
>>>> Feb27-23:00:00 1
>>>> Feb28-01:00:00 1
>>>> Feb28-12:00:00 5
>>>> Feb28-23:00:00 1
>>>> Mar1-01:00:00 1
>>>> Mar1-12:00:00 5
>>>> Mar1-23:00:00 1
>>>> Mar2-01:00:00 1
>>>> Mar2-12:00:00 5
>>>> Mar2-23:00:00 1
>>>> [/test.dat] http://old.nabble.com/file/p27889253/test.png test.png 
>>>> 
>>> 
>>> 
>> 
>> 
> 
> 

-- 
View this message in context: http://old.nabble.com/Gnuplot-is-including-Feb29-even-though-it-is-not-a-leap-year-tp27889253p27906683.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Gnuplot is including Feb29 even though it is not a leap year

[Thomas S. <t.s...@fz...>]

the best way would be to change your parsing script.

adding 10 years inside the 'using' option of the 'plot' command
(like 'using (timecolumn(1)+(2*366+8*365)*24*60*60):2')
doesn't work because the leap year calculation is done before
adding the 10 years.

hardcoding could be done, although i would discourage from
doing this:
in 'time.c' change ZERO_YEAR and adjust JAN_FIRST_WDAY
and SEC_OFFS_SYS accordingly, change 'PROGRAM' in 'setshow.h'
and rename the executable to distinguish it from the official
version.


lozwell wrote:
> 
> Thanks for the reply. Unfortunately I cannot easily set the year in the
> data file, as I am just parsing log files. I could however change my
> parsing script to add a year. Or, is there any way to hard code the year
> in gnuplot? I tried your way, which produced the same results as 2000 is a
> leap year.
> 
> 
> 
> Thomas Sefzick wrote:
>> 
>> change x-axis format to
>> set format x "%b\n%d\n%Y"
>> and you will see that gnuplot assumes its own 'year zero' (=2000) when
>> no information about the year is given.
>> gnuplot can not guess of which year your data are.
>> you should include the year in your data.
>> 
>> 
>> lozwell wrote:
>>> 
>>> Hi, I am having a problem graphing a lot of data using the date/time
>>> over a 2 week period. It looks as though gnuplot is adding in Feb29 to
>>> my x axis with no data points, even though there is no Feb29 as 2010 is
>>> not a leap year. My servers date is the correct year of 2010. This is
>>> causing a gap in my graph to show a large gap. Any ideas how to fix
>>> this? As an example, the following code will also produce the same
>>> results.
>>> 
>>> [test.conf]
>>> set terminal png
>>> set xdata time
>>> set xtic 86400
>>> set timefmt "%b%d-%H:%M:%S"
>>> set format x "%b\n%d"
>>> set yrange [0:10]
>>> set output "test.png"
>>> set grid
>>> plot "test.dat" using 1:2 with linespoints
>>> [/test.conf]
>>> 
>>> [test.dat]
>>> Feb26-01:00:00 1
>>> Feb26-12:00:00 5
>>> Feb26-23:00:00 1
>>> Feb27-01:00:00 1
>>> Feb27-12:00:00 5
>>> Feb27-23:00:00 1
>>> Feb28-01:00:00 1
>>> Feb28-12:00:00 5
>>> Feb28-23:00:00 1
>>> Mar1-01:00:00 1
>>> Mar1-12:00:00 5
>>> Mar1-23:00:00 1
>>> Mar2-01:00:00 1
>>> Mar2-12:00:00 5
>>> Mar2-23:00:00 1
>>> [/test.dat] http://old.nabble.com/file/p27889253/test.png test.png 
>>> 
>> 
>> 
> 
> 

-- 
View this message in context: http://old.nabble.com/Gnuplot-is-including-Feb29-even-though-it-is-not-a-leap-year-tp27889253p27902266.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Line type inside the key

[Thomas S. <t.s...@fz...>]

up to version 4.2.3 the key sample lines are wrong.
(version 4.2.4 i don't have here)
starting with version 4.2.5 the key sample lines are right.
you should upgrade to an actual version of gnuplot.


PiresSLB wrote:
> 
> Hi,
> I'm using the following script:
> 
> set term postscript enhanced color "Times New Roman" 20
> set output "test.eps"
> 
> set style line 3 lt 4 lc 1 lw 1 # red-dotted
> set style line 2 lt 2 lc 2 lw 1 # green-dashed
> set style line 1 lt 1 lc 3 lw 1 # blue-solid
> 
> set key box
> 
> plot "a.txt" with boxes ls 1, "b.txt" with boxes ls 2, "c.txt" with boxes
> ls 3
> 
> where a.txt, b.txt and c.txt are files containing two columns with the bin
> and frequency data for the histogram.
> 
> 
> 
> Thomas Sefzick wrote:
>> 
>> using gnuplot 4.2.2 and the postscript terminal as given, i'm unable to
>> reproduce the wrong lines in the key.
>> 
>> please provide a simple script which produces the reported error.
>> 
>> 
> 
> 

-- 
View this message in context: http://old.nabble.com/Line-type-inside-the-key-tp27871844p27895302.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Gnuplot is including Feb29 even though it is not a leap year

[Thomas S. <t.s...@fz...>]

change x-axis format to
set format x "%b\n%d\n%Y"
and you will see that gnuplot assumes its own 'year zero' (=2000) when
no information about the year is given.
gnuplot can not guess of which year your data are.
you should include the year in your data.


lozwell wrote:
> 
> Hi, I am having a problem graphing a lot of data using the date/time over
> a 2 week period. It looks as though gnuplot is adding in Feb29 to my x
> axis with no data points, even though there is no Feb29 as 2010 is not a
> leap year. My servers date is the correct year of 2010. This is causing a
> gap in my graph to show a large gap. Any ideas how to fix this? As an
> example, the following code will also produce the same results.
> 
> [test.conf]
> set terminal png
> set xdata time
> set xtic 86400
> set timefmt "%b%d-%H:%M:%S"
> set format x "%b\n%d"
> set yrange [0:10]
> set output "test.png"
> set grid
> plot "test.dat" using 1:2 with linespoints
> [/test.conf]
> 
> [test.dat]
> Feb26-01:00:00 1
> Feb26-12:00:00 5
> Feb26-23:00:00 1
> Feb27-01:00:00 1
> Feb27-12:00:00 5
> Feb27-23:00:00 1
> Feb28-01:00:00 1
> Feb28-12:00:00 5
> Feb28-23:00:00 1
> Mar1-01:00:00 1
> Mar1-12:00:00 5
> Mar1-23:00:00 1
> Mar2-01:00:00 1
> Mar2-12:00:00 5
> Mar2-23:00:00 1
> [/test.dat] http://old.nabble.com/file/p27889253/test.png test.png 
> 

-- 
View this message in context: http://old.nabble.com/Gnuplot-is-including-Feb29-even-though-it-is-not-a-leap-year-tp27889253p27895111.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] How to dynamically specify color for drawing circle in Gnuplot4.3

[v_2e <v_...@uk...>]

  Hello!

TonyM wrote:
> 
> I'm having the same problem.
> Can you please provide a working example of both the data and the code
> used to run.
> 
  I can provide my working example. It is very simple.
I have a test_map.txt file with contents like this:
0 0 5 7
-90 0 2 4
90 0 2 4
-120 0 1 3
The columns mean x-coordinate, y-coordinate, radius and a color
respectively.
  And I can easily plot the circles from such file just by using this
command line:
$ gnuplot -persist -e 'plot "test_map.txt" using 1:2:3:4 with circles lc
variable'
  Or if I would like to have filled circles, I can use:
$ gnuplot -persist -e 'plot "test_map.txt" using 1:2:3:4 with circles fill
solid lc variable'
  And that is all! :)
  Hope, it will help you!

-- 
View this message in context: http://old.nabble.com/How-to-dynamically-specify-color-for-drawing-circle-in-Gnuplot4.3-tp21376845p27892177.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Line type inside the key

[Thomas S. <t.s...@fz...>]

using gnuplot 4.2.2 and the postscript terminal as given, i'm unable to
reproduce the wrong lines in the key.

please provide a simple script which produces the reported error.


Hi,
I'm using GNUPLOT 	Version 4.2 patchlevel 2 with set term postscript
enhanced color "Times New Roman" 20


Thomas Sefzick wrote:
> 
> with gnuplot version are you using?
> with terminal are you using?
> 
> 

-- 
View this message in context: http://old.nabble.com/Line-type-inside-the-key-tp27871844p27887576.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Line type inside the key

[Thomas S. <t.s...@fz...>]

with gnuplot version are you using?
with terminal are you using?


PiresSLB wrote:
> 
> Hi,
> I am plotting several histograms and to do that nicely I use the style
> line command to specify both the colour and the line type. I use the
> following command to get want I want:
> 
> set style line 3 lt 4 lc 1 lw 1 # red-dotted
> set style line 2 lt 2 lc 2 lw 1 # green-dashed
> set style line 1 lt 1 lc 3 lw 1 # blue-solid
> 
> This works fine the only problem I have is that I want the key to the plot
> to be red-dotted, green-dashed and blue-solid. Instead of this I have the
> key with red filled, green filled and blue filled. Basically the line
> types that I use in the plot don't match the key. 
> 
> Why does this happen? If I use the default style lines this doesn't
> happen, but If i want to specify the colour and line type of my plot the
> key doesn't match anymore.
> 
> Cheers
> 

-- 
View this message in context: http://old.nabble.com/Line-type-inside-the-key-tp27871844p27874922.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Adding the value to the line

[Thomas S. <t.s...@fz...>]

set xdata time
set timefmt "%H:%M:%S"
set offsets 0,0,1,0
plot 'datafilename' using 1:2 with linespoints, '' using 1:($2>5?$2:0/0):2
with labels offset 0,1


lozwell wrote:
> 
> I am graphing some data that show large spikes. What I would like to do is
> add the actual value to the top of the spikes or half way up the spike. Is
> there a way to do this? For instance, with the following data, I'd like to
> add the y value to only the entries that are over 5. Any help
> appreciated!!
> 
> 01:00:00	1
> 02:00:00	1
> 03:00:00	7
> 04:00:00	2
> 05:00:00	3
> 06:00:00	10
> 07:00:00	1
> 08:00:00	4
> 09:00:00	2
> 10:00:00	9
> 11:00:00	3
> 12:00:00	2
> 

-- 
View this message in context: http://old.nabble.com/Adding-the-value-to-the-line-tp27869657p27874846.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

[Gnuplot-info] transparency

[Tamas H. <ho...@gm...>]

Is it possible to have transparency effect on lines.
If yes, is it possible to change transparency dynamically?

[Gnuplot-info] error bar color in histogram

[Zhong R. <re...@uc...>]

Hi,

I would like to specify colors for my histogram, but the error
bars do not follow my specification.  How do I make each error
bar same color as its histogram column?

Zhong

[Gnuplot-info] variable color in histogram

[Zhong R. <re...@uc...>]

Hi,

Is there a way to use variable rgb colors for histogram bars?
 I tried to combine examples from histogram and variable
colors, but got error that states:

Too many columns in using specification

The plot line looks like this:

plot newhistogram '...', '-' using 2:xtic(1):6 lc rgb
variable, ...

I hoped that rgb variable would pick up the last column :6,
but obviously it doesn't like that.  Is there a way to do
this?  Thanks.

Zhong

Re: [Gnuplot-info] Fourier series fit works in matlab but fails in gnuplot :(

[Thomas S. <t.s...@fz...>]

start the fit with only a few parameters (the important ones),
and then increase the number of parameters:

- set a1,...,a8 and b1,...,b8 to 0.0, w to 1.0
- fit a0 first: fit f8(x) 'datafilename' via a0
- then set a1,b1,w by hand and compare data and function:
   plot 'datafilename' with lines, f8(x)
- once you've found the right values (especially for 'w'), include
   more parameters in your fit:
   a2=1;b2=1;fit f8(x) 'datafilename' via a0,a1,b1,a2,b2,w
   plot 'datafilename' with lines, f8(x)
   a3=1;b3=1;fit f8(x) 'datafilename' via a0,a1,b1,a2,b2,a3,b3,w
   plot 'datafilename' with lines, f8(x)
   a.s.o.


mishkind wrote:
> 
> Tatsuro:
> The matlab fit I was doing also used Marquardt-Levenberg algorithm,
> but like you said, perhaps they have a good initial parameter
> determining engine.
> The extra details in the subtopics of fit were very useful as I missed
> them the first time around.
> 
> Hans-Bernhard:
> Thanks for taking a look at my data. I've tried normalizing my
> parameters as there was around 4 orders of magnitude difference, but
> that hasn't helped.
> 
> I guess I have to find a good way to initialize my parameters.
> Anybody with some fourier series theory that could lend a hand? The
> only tricks I know are:
> - a0 being the DC term
> - (a1,a2...)=0 for an odd function and (b1,b2..)=0 for an even function
> - higher frequencies= sharp changes in the data, lower= smooth changes.
> 
> What I really could use a hand with is just determining the sign of
> each parameter?
> 
> thanks,
> mishkin
> 
> ------------------------------------------------------------------------------
> Download Intel® Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 
> 

-- 
View this message in context: http://old.nabble.com/Fourier-series-fit-works-in-matlab-but-fails-in-gnuplot-%3A%28-tp27792951p27832575.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Saving mutiple plots as mutliple jpegs simultaenously

[Thomas S. <t.s...@fz...>]

loop.plt:

i=i+1
set output "test".i.".plt"
show output
if (i<=n) reread

at the gnuplot prompt:

i=0
n=10
load 'loop.plt'


Hi,

I have taken a look at the reread command. Seems useful but looks like I
can`t
use it for what I need to do: reread can increament some variable
but what I need to do is to increament the filename
e.g
set output "test.png"

then

set output "test2.png"
set output "test3.png"

Doesn't look like I can use reread, or can I?



Thomas Sefzick wrote:
> 
> have a look at the 'reread' command:
> http://www.gnuplot.info/docs/gnuplot.html#x1-15300040
> 
> 
> 

-- 
View this message in context: http://old.nabble.com/Saving-mutiple-plots-as-mutliple-jpegs-simultaenously-tp27811835p27823329.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Saving mutiple plots as mutliple jpegs simultaenously

[Thomas S. <t.s...@fz...>]

have a look at the 'reread' command:
http://www.gnuplot.info/docs/gnuplot.html#x1-15300040


James00007 wrote:
> 
> Hi,
> 
> I`m a new user of gnuplot
> 
> I`ve figured out how to save a plot as a jpeg.
> 
> I now need to figure out how to make plots of multiple sequentially named
> data files and save the
> individual plots as jpegs. 
> 
> I can do it the manual way repeating the command so many times but I　was
> wondering if anyone
> could show me a faster was with a script or something 
> 

-- 
View this message in context: http://old.nabble.com/Saving-mutiple-plots-as-mutliple-jpegs-simultaenously-tp27811835p27819659.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.