gnuplot-info — For questions and discussion about gnuplot

[Gnuplot-info] Problems with the display

[Mauricio F. <mau...@gm...>]

Hello:

I am trying to use gnuplot in cygwin, in my windows, and I am currently
having a problem. I ssh into cygwin in the following way:

ssh -X user@address

and then, I type:

gnuplot.

And I get a command prompt that looks like this:

gnuplot>

So I type in:

plot x

And the computer does nothing, it just shows a new command prompt line like
this:

gnuplot>

So I basically came up with the conclusion that my cygwin does not have a
display screen or something, is that right? Did I make any mistake in my use
of gnuplot? Do you know how can I get the display to be shown in my
computer?

Thanks,

Mauricio Flores
The University of Texas at Austin

[Gnuplot-info] datafile separator

[János L. <jan...@ya...>]

Hi,

I am using tab - "\t" - in a datafile as field separator.  I have 33 fields in a row.  I set in the plotting script the datafile separator explicitly.

set datafile separator "\t"

I expected when gnuplot sees TWO tabs after each other it treats the nonexistent data in-between these two tabs as a field with no data.  Unfortunately it is not the case.  It swallows bot tabs as one tab and let say if it is the 28th field that has no data, it plots the value of the 29th field of that row.

What can I do to change this behavior ?  Changing from tab to vertical bar, or anything else, is not the solution, because the double vertical bar would be treated as jut one, the same way as it is done for the tabs.

HELP !!! :-)

Here are two lines from the datafile to illustrate:

Oct 26, 2009 11:04:37	16257	2081	2136	1809	1651	3011	3102	2698	6961	4158	2176	6028	2262	3604	3766	4535	4096	3916	8443	4423	4116	5199	2950	5368	9619		0	85353	107744	2625	87907
Oct 26, 2009 12:04:26	26285	1603	4534	1237	1151	1584	2147	2193	4749	3659	7776	5746	1762	2951	3301	3592	3141	3208	3539	7886	3609	3848	2522	4485	8374	0	960	78954	129196	2562	115163

In the first record between 9619 and 0 there are TWO tabs.  On the second record in the same position between 8374 and 960 there is a field value 0.  I want gnuplot in the case of the first record NOT to plot 0 as a value for field 27.  Right now it does, and that 0 belongs to field 28.

Thanks ahead,

János

Re: [Gnuplot-info] $field is not recognized

[János L. <jan...@ya...>]

On Jun 23, 2010, at 9:23 AM, Hans-Bernhard Bröker wrote:

> Am 17.06.2010 18:57, schrieb Janos Lobb:
>> gnuplot>  call '/Volumes/Data/PROJECTS/Sysmon/gnuplot_cmds/context.gp'
> 
> That makes no sense. 'call' is for loading gnuplot scripts that take additional arguments.  Yours doesn't, so you should simply 'load' it instead.
> 
>> As you can see it lost the $field variables.
> 
> That's because in a 'call'ed script, $<number> means something different.  See "help call".

Hi Hans-Bernhard,

I use call, because later on I will pass variables :-)

In the meantime - reading the Fine Manual - I realized that $x is just a shorthand for column(x).  Doing that replacement my problem disappeared.

Thanks a lot,

János

Re: [Gnuplot-info] Using title columnheader

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

the black sample lines are due to a bug which is in version 4.4.0
the development version (4.5) works properly.


Frazer Williams wrote:
> 
> 
> Thomas Sefzick <t.sefzick <at> fz-juelich.de> writes:
> 
> 
>> plot for [i=0:1] "graph.dat" using 1:2:-2 index i lc variable title
>> columnheader
> 
> 
> Thanks for the reply.  Your suggestion almost works.  When I use it, I get
> a
> plot with multiple lines, each with a different color.  I also get a key
> with
> multiple titles, one for each line.  The problem is that all the sample
> lines in
> the key are solid black, so it is difficult to know which line on the plot
> belongs with each title.  Is there something else I should be doing?
> 
> Also, a minor problem is that it would be nice to be able to determine the
> value
> of n from the file, something like
> 
>   plot for [i=0:n]
> 
> where n is determined from "graph.dat".  A work-around is
>   
>   plot for [i=0:20]
> 
> where I assume the number of lines to be plotted is < 20.  That gives some
> warnings, but works otherwise.
> 
> Thank you for your time.
> 
> Frazer
> 
> P.S.  I submitted this followup several days ago, but it didn't appear, so
> I'm
> reposting.  I apologize if it becomes a multiple post.
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate 
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
> lucky parental unit.  See the prize list and enter to win: 
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 
> 

-- 
View this message in context: http://old.nabble.com/Using-title-columnheader-tp28907621p28973534.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] $field is not recognized

[Hans-Bernhard B. <HBB...@t-...>]

Am 17.06.2010 18:57, schrieb Janos Lobb:
> gnuplot>  call '/Volumes/Data/PROJECTS/Sysmon/gnuplot_cmds/context.gp'

That makes no sense. 'call' is for loading gnuplot scripts that take 
additional arguments.  Yours doesn't, so you should simply 'load' it 
instead.

> As you can see it lost the $field variables.

That's because in a 'call'ed script, $<number> means something 
different.  See "help call".

[Gnuplot-info] Data matrix plot with not valid grid points

[Seoane A. T. <ts...@e-...>]

Hello!

I have a problem, long to explain, but not complex to understand:

I am trying to plot with image a data matrix with the following
properties:
- The valid data goes from 0 to 1.
- There exist invalid data, which has a z value of 2.
- The palette range will be [0:0.01]

I want 29 colors to be used in the palette, and so, I use this script:

set cbrange [0:0.01]
set palette model RGB defined ( 0 "#8b0000", 1 "#ff0000", 2 "#ffa500", 3
"#ffff00", 4 "#00ff00", 5 "#00ffff",  6 "#0000ff", 7 "#00008b" )
set pal maxcolors 29
set view map
splot 'matrix.txt' with image

With this script, all the valid data from 0.01 to 1 will have the last
color of the palette, which is OK, but the invalid data will have that
color too, but I would like those data points to have another color. Is
there any simple way to achieve this, like skipping the points with
certain z values?

As I could not find a simple solution, I have tried the next more or
less complex approach which "almost" works:
- Redefine the palette range to include the invalid data.
- Redefine the palette with its real range values to have the same
colors than before for the same values.

set cbrange [0:2]
set palette model RGB defined ( a "#8b0000", b "#8b0000", b "#a80000", c
"#a80000", c "#c50000", d "#c50000", d "#e20000", e "#e20000", e
"#ff0000", f "#ff0000", f "#ff2900", g "#ff2900", g "#ff5300", h
"#ff5300", h "#ff7c00", i "#ff7c00", i "#ffa500", j "#ffa500", j
"#ffbc00", k "#ffbc00", k "#ffd200", l "#ffd200", l "#ffe900", m
"#ffe900", m "#ffff00", n "#ffff00", n "#bfff00", o "#bfff00", o
"#7fff00", p "#7fff00", p "#40ff00", q "#40ff00", q "#00ff00", r
"#00ff00", r "#00ff40", s "#00ff40", s "#00ff80", t "#00ff80", t
"#00ffbf", u "#00ffbf", u "#00ffff", v "#00ffff", v "#00bfff", w
"#00bfff", w "#0080ff", x "#0080ff", x "#0040ff", y "#0040ff", y
"#0000ff", z "#0000ff", z "#0000e2", a1 "#0000e2", a1 "#0000c5", b1
"#0000c5", b1 "#0000a8", c1 "#0000a8", c1 "#00008b",1 .1 "#00008b", 1.1
"#bebebe", 2 "#bebebe" )
set view map
splot 'matrix.txt' with image

Where the parameters "a" to "c1" has the next value:
0. + 0.01 * i / 29.
i=0, 1, 2, ...

The problem with this is that not all colors are used, but I have found
that if the palette range is smaller, it works OK, so maybe is a gnuplot
bug or limitation. Anyway, I cannot do the range smaller, as there would
be valid data with the color of invalid data (in the example there is no
valid data from 0.01 to 1, so it is perfect, but not always will be this
way).
This would be the script for the limited range that works:

set cbrange [0:.05]
set palette model RGB defined ( a "#8b0000", b "#8b0000", b "#a80000", c
"#a80000", c "#c50000", d "#c50000", d "#e20000", e "#e20000", e
"#ff0000", f "#ff0000", f "#ff2900", g "#ff2900", g "#ff5300", h
"#ff5300", h "#ff7c00", i "#ff7c00", i "#ffa500", j "#ffa500", j
"#ffbc00", k "#ffbc00", k "#ffd200", l "#ffd200", l "#ffe900", m
"#ffe900", m "#ffff00", n "#ffff00", n "#bfff00", o "#bfff00", o
"#7fff00", p "#7fff00", p "#40ff00", q "#40ff00", q "#00ff00", r
"#00ff00", r "#00ff40", s "#00ff40", s "#00ff80", t "#00ff80", t
"#00ffbf", u "#00ffbf", u "#00ffff", v "#00ffff", v "#00bfff", w
"#00bfff", w "#0080ff", x "#0080ff", x "#0040ff", y "#0040ff", y
"#0000ff", z "#0000ff", z "#0000e2", a1 "#0000e2", a1 "#0000c5", b1
"#0000c5", b1 "#0000a8", c1 "#0000a8", c1 "#00008b", .03 "#00008b", .03
"#bebebe", .05 "#bebebe" )
set view map
splot 'matrix.txt' with image

I would be very grateful if anyone could give me a solution!

I have attached 3 images for the 3 scripts so you can see the problem.

Thanks!

Teo Seoane




--------------------------------------------------------------------------
Advertencia / Confidentiality Notice

Este envio es confidencial y esta destinado unicamente a la persona a la que ha sido enviado. Puede contener informacion privada y confidencial. Si usted no es el destinatario al que ha sido remitida, no puede copiarla, distribuirla ni emprender con ella ningun tipo de accion. Si cree que lo ha recibido por error, por favor, notifiquelo al remitente.

This transmission is confidential and intended solely for the person to whom it is addressed. It may contain privileged and confidential information. If you are not the intended recipient, you should not copy, distribute or take any action in reliance on it. If you believe that you have received this transmission in error, please notify the sender.

Aeropuertos Espanoles y Navegacion Aerea

Re: [Gnuplot-info] Using title columnheader

[Frazer <pf...@mo...>]

Thomas Sefzick <t.sefzick <at> fz-juelich.de> writes:


> plot for [i=0:1] "graph.dat" using 1:2:-2 index i lc variable title
> columnheader


Thanks for the reply.  Your suggestion almost works.  When I use it, I get a
plot with multiple lines, each with a different color.  I also get a key with
multiple titles, one for each line.  The problem is that all the sample lines in
the key are solid black, so it is difficult to know which line on the plot
belongs with each title.  Is there something else I should be doing?

Also, a minor problem is that it would be nice to be able to determine the value
of n from the file, something like

  plot for [i=0:n]

where n is determined from "graph.dat".  A work-around is
  
  plot for [i=0:20]

where I assume the number of lines to be plotted is < 20.  That gives some
warnings, but works otherwise.

Thank you for your time.

Frazer

P.S.  I submitted this followup several days ago, but it didn't appear, so I'm
reposting.  I apologize if it becomes a multiple post.

Re: [Gnuplot-info] how to type the times symbol in eps figures?

[thorneliu <liu...@co...>]

Hans-Bernhard Bröker-2 wrote:
> 
> Am 20.06.2010 14:34, schrieb thorneliu:
> 
>> I am using gnuplot and I want to label the times symbol in my figure (you
>> know , just like the $\times$ symbol in TeX)
>> what is the code for that symbol ?
> 
> That depends on what output format you're trying to generate (--> help 
> terminal).  In some of them it actually is $\times$, because there are
> TeX.
> 
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate 
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
> lucky parental unit.  See the prize list and enter to win: 
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 
> 
thank you...I am using the terminal " post enh " (eps format)
just now I refered to the documents on
http://t16web.lanl.gov/Kawano/gnuplot/tics-e.html
And I got the answer to my question 
{\Symbol \264} works
-- 
View this message in context: http://old.nabble.com/how-to-type-the-times-symbol-in-eps-figures--tp28940563p28940836.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] how to type the times symbol in eps figures?

[Hans-Bernhard B. <HBB...@t-...>]

Am 20.06.2010 14:34, schrieb thorneliu:

> I am using gnuplot and I want to label the times symbol in my figure (you
> know , just like the $\times$ symbol in TeX)
> what is the code for that symbol ?

That depends on what output format you're trying to generate (--> help 
terminal).  In some of them it actually is $\times$, because there are TeX.

[Gnuplot-info] how to type the times symbol in eps figures?

[thorneliu <liu...@co...>]

Hello, everyone!
I am using gnuplot and I want to label the times symbol in my figure (you
know , just like the $\times$ symbol in TeX)
what is the code for that symbol ? 
thanks in advance.
-- 
View this message in context: http://old.nabble.com/how-to-type-the-times-symbol-in-eps-figures--tp28940563p28940563.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Why would gnuplot ignore my first data point in a file?

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

if you define

set key autotitles columnhead

the first data line is used for setting the title, so it's
not used for plotting.
you would have seen this if you would have let gnuplot
set the title automatically, instead you specified an explicit
title as option for the plot command.


Christopher D. Carothers wrote:
> 
> When I remove the copy of the first line, it disappears ..
> 
> What you have it my hack to get it to appear .. just copy
> the first line twice in the data file ..
> 
> It appears to have the same behavior on gnuplot 4.2.5 ...
> 
> Any ideas??
> 
> Thanks,
> Chris
> 
> 2010/6/15 Hans-Bernhard Bröker <HBB...@t-...>
> 
>> Christopher D. Carothers wrote:
>>
>>> Hi  -- I've been a long-time user of gnuplot, but am baffled at
>>> why it's consistently ignoring the first line from all my data input
>>> files?
>>>
>>> I am using the current release version, 4.4.0.
>>>
>>> My data file is:
>>>
>>>
>>> 50534   0.01   2048   4   2   1          42331740.77
>>> 50534   0.01   2048   4   2   1          42331740.77
>>>
>>
>> How can you possibly have managed to tell whether or not gnuplot has
>> ignored the first data point from this file, given that it's exactly the
>> same as the second?
>>
>>
>>
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate 
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
> lucky parental unit.  See the prize list and enter to win: 
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 
> 

-- 
View this message in context: http://old.nabble.com/Why-would-gnuplot-ignore-my-first-data-point-in-a-file--tp28894854p28935654.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] rowstack with 2 columns per entry

[mkn86 <m.k...@gm...>]

Thx a lot!!!! Didn't expect the solutation that simple :)
-- 
View this message in context: http://old.nabble.com/rowstack-with-2-columns-per-entry-tp28933153p28935646.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] rowstack with 2 columns per entry

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

start a new histogram series for the 3rd column:

reset
set title "Query 1 - BSBM"
set style data histogram
set boxwidth 0.3 absolute
set style histogram rowstacked
set style fill solid 1.0 border lt -1
set ylabel "Number of disk page reads"
set xtics offset first -0.15,0
plot "query1.dat" using 2 t "sequential read", "" using 3 t "random read",
newhistogram at 0.3, "" using 4:xtic(1) t "our approach"


mkn86 wrote:
> 
> Hi!
> 
> I need to plot a bar chart as a rowstacked histogram.
> The first picture "histo1.jpg" shows the diagram without rowstacking.
>  http://old.nabble.com/file/p28933153/histo1.jpg 
> My problem now is that I want the red bar on top of the green bar
> (sequential read + random read). So in sum, I will have for each data set
> (1mio, 2mio, ..., 31mio) two bar charts, namely a rowstacked one (green +
> red) and the blue one.
> Therefore I run this commands:
> set title "Query 1 - BSBM"
> set style data histogram
> set boxwidth 0.75 absolute
> set style histogram rowstacked
> set style fill solid 1.0 border lt -1
> set yrange [0:300]
> set ylabel "Number of disk page reads"
> plot "query1.dat" using 2 t "sequential read", "" using 3 t "random read",
> "" using 4:xtic(1) t "our approach"
> 
> which delivers me this:
>  http://old.nabble.com/file/p28933153/histo2.jpg 
> Unfortunately the blue bar (our approach) is also stacked up. Does someone
> know how I can have two bars per dataset? 
> Thanks a lot in advance for any help!
> 
> Cheers
> 
> P.S.
> the dat file looks like this:
> #	sequential page	random page	our model
> 1mio		6	67	38
> 2mio		6	102	54
> 5mio		8	100	54
> 10mio	4	104	54
> 31mio	12	120	54
> 

-- 
View this message in context: http://old.nabble.com/rowstack-with-2-columns-per-entry-tp28933153p28933765.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

[Gnuplot-info] rowstack with 2 columns per entry

[mkn86 <m.k...@gm...>]

Hi!

I need to plot a bar chart as a rowstacked histogram.
The first picture "histo1.jpg" shows the diagram without rowstacking.
http://old.nabble.com/file/p28933153/histo1.jpg 
My problem now is that I want the red bar on top of the green bar
(sequential read + random read). So in sum, I will have for each data set
(1mio, 2mio, ..., 31mio) two bar charts, namely a rowstacked one (green +
red) and the blue one.
Therefore I run this commands:
set title "Query 1 - BSBM"
set style data histogram
set boxwidth 0.75 absolute
set style histogram rowstacked
set style fill solid 1.0 border lt -1
set yrange [0:300]
set ylabel "Number of disk page reads"
plot "query1.dat" using 2 t "sequential read", "" using 3 t "random read",
"" using 4:xtic(1) t "our approach"

which delivers me this:
http://old.nabble.com/file/p28933153/histo2.jpg 
Unfortunately the blue bar (our approach) is also stacked up. Does someone
know how I can have two bars per dataset? 
Thanks a lot in advance for any help!

Cheers

P.S.
the dat file looks like this:
#	sequential page	random page	our model
1mio		6	67	38
2mio		6	102	54
5mio		8	100	54
10mio	4	104	54
31mio	12	120	54
-- 
View this message in context: http://old.nabble.com/rowstack-with-2-columns-per-entry-tp28933153p28933153.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Why would gnuplot ignore my first data point in a file?

[Philipp K. J. <gn...@be...>]

What is the smallest file that shows this behavior?

For instance, if you have the file containing ONLY:

1
2
3

does it gobble up the 1 and the 3?

I wonder whether your problem is something weird
having to do w/ file formats - such as taking a file
from one platform and trying to read it on another.

Best,

		Ph.


On Tuesday 15 June 2010 12:14:45 pm Christopher D. Carothers wrote:
> When I remove the copy of the first line, it disappears ..
>
> What you have it my hack to get it to appear .. just copy
> the first line twice in the data file ..
>
> It appears to have the same behavior on gnuplot 4.2.5 ...
>
> Any ideas??
>
> Thanks,
> Chris
>
> 2010/6/15 Hans-Bernhard Bröker <HBB...@t-...>
>
> > Christopher D. Carothers wrote:
> >> Hi  -- I've been a long-time user of gnuplot, but am baffled at
> >> why it's consistently ignoring the first line from all my data input
> >> files?
> >>
> >> I am using the current release version, 4.4.0.
> >>
> >> My data file is:
> >>
> >>
> >> 50534   0.01   2048   4   2   1          42331740.77
> >> 50534   0.01   2048   4   2   1          42331740.77
> >
> > How can you possibly have managed to tell whether or not gnuplot has
> > ignored the first data point from this file, given that it's exactly the
> > same as the second?
>
> ---------------------------------------------------------------------------
>--- ThinkGeek and WIRED's GeekDad team up for the Ultimate
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
> lucky parental unit.  See the prize list and enter to win:
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info

Re: [Gnuplot-info] Using title columnheader

[Frazer <pf...@mo...>]

Thomas Sefzick <t.sefzick <at> fz-juelich.de> writes:


> plot for [i=0:1] "graph.dat" using 1:2:-2 index i lc variable title
> columnheader


Thanks.  That almost works.  I get the plots I want, with the graph lines in
different colors, but the sample lines in the keys are all solid black, so the
keys don't help a lot.  Is there something else I should do?

On related topic, it would also be nice if there were some way to read in the
number of data sets from the input file, so I could write something like

plot for [i=0:n]

Is there any way to do that?  It appears I can just write

plot for [i=0:20]

where 20 is assumed larger than the number of data sets I'll ever have.  That
gets me some warnings, but otherwise seems to work.

Thanks for the response and for your time.

Frazer 

Re: [Gnuplot-info] Using title columnheader

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

plot for [i=0:1] "graph.dat" using 1:2:-2 index i lc variable title
columnheader


Frazer Williams wrote:
> 
> Hi,
> 
> I generate data files consisting of x,y data for a number of different
> times.
> I would like gnuplot to plot the data for each of the times, along with a
> title
> that is something like "t=0.1".  So far I have been unsuccessful.  I would
> appreciate suggestions.  I can write the data file in any format that
> works.
> 
> I am putting the data for each time sequentially in the file, with two
> blank
> lines between data sets.  I am putting a line like
> 
> TIME "t=0.1"
> 
> at the start of each set.  In other words, the file looks like
> 
> TIME "t=0.1"
> 1  10
> 2  20
> 3  40
> 
> 
> TIME "t=0.2"
> 1  20
> 2  30
> 3  53
> 
> and so forth.
> 
> I have tried to plot it with
> 
> plot "graph.dat" using 1:2:((column(-2)) lc variable title columnheader
> 
> This gives me the first (t=0.1) plot, with the correct title, but the
> remainder
> of the sets are plotted but have no title.  I get a similar result with
> 
> set key autotitle columnhead
> plot "graph.dat" using 1:2:((column(02)) lc variable
> 
> I am using gnuplot 4.4, patchlevel 0.
> 
> It seems this ought to be possible.  Any help would be appreciated.
> 
> Thanks.
> 
> Frazer
> 
> 
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate 
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
> lucky parental unit.  See the prize list and enter to win: 
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 
> 

-- 
View this message in context: http://old.nabble.com/Using-title-columnheader-tp28907621p28922457.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

[Gnuplot-info] $field is not recognized

[Janos L. <jan...@ya...>]

Hi,

I am on a Mac with OS 10.6.3 and using gnuplot  4.4 patchlevel 0.
This is the content of my context.gp file:

set xdata time
set timefmt "%b %d, %Y %H:%M:S"
set format x "%m\n%y"
set title "Context Switches"
set logscale y
set xlabel "DateTime [month/year]"
set ylabel "Context Switches [#/hour]"
set terminal postscript eps color
set output '/Volumes/Data/PROJECTS/Sysmon/graphs/context.ps'
plot [][] "/Volumes/Data/PROJECTS/Sysmon/outs/context.out" 
     using 1:5 title "Voluntary Yields" smooth bezier with lines,
 "" using 1:($6 > 0 ? $6 : 1) title "Cache Misses" smooth bezier with lines,
 "" using 1:($7+1) title "System Disk Writes" smooth bezier with lines,
 "" using 1:($8+1) title "I/O Paceing" smooth bezier with lines,
 "" using 1:($9+1) title "Logical Lock Contention" smooth bezier with lines
set terminal x11
set output
set xdata
set timefmt
set format x
unset logscale y
set xlabel
set ylabel

I load this with :
gnuplot> call '/Volumes/Data/PROJECTS/Sysmon/gnuplot_cmds/context.gp'

gnuplot complains this way:
gnuplot> plot [][] "/Volumes/Data/PROJECTS/Sysmon/outs/context.out" 
using 1:5 title "Voluntary Yields" smooth bezier with lines,
 "" using 1:( > 0 ?  : 1) title "Cache Search Misses" smooth bezier with lines,
 "" using 1:(+1) title "System Disk Writes" smooth bezier with lines,
 "" using 1:(+1) title "I/O Paceing" smooth bezier with lines,
 "" using 1:(+1) title "Logical Lock Contention" smooth bezier with 
lines
                                                                                                                                              ^
"/Volumes/Data/PROJECTS/Sysmon/gnuplot_cmds/context.gp",
 line 11: invalid expression 

As you can see it lost the $field variables.

If I do it interactively and do not set the terminal to postscript and the 
output to the file, the gnuplot does not complain and I see the graphic.

What am I doing wrong ? 

Thanks ahead,

János

Re: [Gnuplot-info] Redhat EL 5.3 no x11?

[Hans-Bernhard B. <HBB...@t-...>]

Xia Qu wrote:

> Today I updated gnuplot in my computer which runs redhat EL 5.3 from version
> 4.0 to 4.4. But after the installation, I find out that now gnuplot has no
> x11 terminal. 

What kind of installation was that?  I'll venture a guess that you 
compiled your own gnuplot.  And that you didn't have the development 
packages for X11 installed.  And you didn't pay particularly close 
attention to the output of the ./configure command, which tried to tell 
about the lacking X11 development support.

> I don't know how to fix this. 

Install the missing parts before building gnuplot.  Or maybe install a 
binary package of gnuplot by your distributor.

[Gnuplot-info] Using title columnheader

[Frazer <pf...@mo...>]

Hi,

I generate data files consisting of x,y data for a number of different times.
I would like gnuplot to plot the data for each of the times, along with a title
that is something like "t=0.1".  So far I have been unsuccessful.  I would
appreciate suggestions.  I can write the data file in any format that works.

I am putting the data for each time sequentially in the file, with two blank
lines between data sets.  I am putting a line like

TIME "t=0.1"

at the start of each set.  In other words, the file looks like

TIME "t=0.1"
1  10
2  20
3  40


TIME "t=0.2"
1  20
2  30
3  53

and so forth.

I have tried to plot it with

plot "graph.dat" using 1:2:((column(-2)) lc variable title columnheader

This gives me the first (t=0.1) plot, with the correct title, but the remainder
of the sets are plotted but have no title.  I get a similar result with

set key autotitle columnhead
plot "graph.dat" using 1:2:((column(02)) lc variable

I am using gnuplot 4.4, patchlevel 0.

It seems this ought to be possible.  Any help would be appreciated.

Thanks.

Frazer

Re: [Gnuplot-info] data file output for bezier smoothing

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

set table


ranlot wrote:
> 
> How do you tell gnuplot to output the x,y coordinates of the curves it
> draws when asking for a bezier smoothing?  One may need to use this raw
> data instead of just a direct visualization of the curve.
> 

-- 
View this message in context: http://old.nabble.com/data-file-output-for-bezier-smoothing-tp28901357p28901412.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

[Gnuplot-info] data file output for bezier smoothing

[ranlot <ran...@ho...>]

How do you tell gnuplot to output the x,y coordinates of the curves it draws
when asking for a bezier smoothing?  One may need to use this raw data
instead of just a direct visualization of the curve.
-- 
View this message in context: http://old.nabble.com/data-file-output-for-bezier-smoothing-tp28901357p28901357.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

[Gnuplot-info] Redhat EL 5.3 no x11?

[Xia Qu <qux...@gm...>]

Hi,

Today I updated gnuplot in my computer which runs redhat EL 5.3 from version
4.0 to 4.4. But after the installation, I find out that now gnuplot has no
x11 terminal. I don't know how to fix this. And I've installed all different
versions, even back to v4.0, but x11 just not there. Anyone knows how to
solve this?

Thanks

Re: [Gnuplot-info] Why would gnuplot ignore my first data point in a file?

[Christopher D. C. <ch...@cs...>]

When I remove the copy of the first line, it disappears ..

What you have it my hack to get it to appear .. just copy
the first line twice in the data file ..

It appears to have the same behavior on gnuplot 4.2.5 ...

Any ideas??

Thanks,
Chris

2010/6/15 Hans-Bernhard Bröker <HBB...@t-...>

> Christopher D. Carothers wrote:
>
>> Hi  -- I've been a long-time user of gnuplot, but am baffled at
>> why it's consistently ignoring the first line from all my data input
>> files?
>>
>> I am using the current release version, 4.4.0.
>>
>> My data file is:
>>
>>
>> 50534   0.01   2048   4   2   1          42331740.77
>> 50534   0.01   2048   4   2   1          42331740.77
>>
>
> How can you possibly have managed to tell whether or not gnuplot has
> ignored the first data point from this file, given that it's exactly the
> same as the second?
>
>
>

Re: [Gnuplot-info] Why would gnuplot ignore my first data point in a file?

[Hans-Bernhard B. <HBB...@t-...>]

Christopher D. Carothers wrote:
> Hi  -- I've been a long-time user of gnuplot, but am baffled at
> why it's consistently ignoring the first line from all my data input files?
> 
> I am using the current release version, 4.4.0.
> 
> My data file is:
> 
> 
> 50534   0.01   2048   4   2   1          42331740.77
> 50534   0.01   2048   4   2   1          42331740.77

How can you possibly have managed to tell whether or not gnuplot has 
ignored the first data point from this file, given that it's exactly the 
same as the second?