Hi Steve,

Yes this does seem to do what I need, many thanks.

kevin

--- In ploticus@yahoogroups.com, Stephen Grubb <Stephen.Grubb@...> wrote:
>
> Hello Kevin,
>
> I tried it with mode=line and removed the xrange argument, and my
gif result is attached.
>
> It looks like online time is very short intervals, while offline
time is quite a bit more, so the result seems reasonable to me.
>
> Feel free to follow up if you wish,
>
> Steve
> ________________________________________

Hi, Sir:

We want to use the ploticus to add charts into our Mediawiki-powered
website: http://www.cn0km.com/ through an extension called
Easytimeline. I tried for a while but still cannot get it. Whether
this executable pl tool is installed by the site host? We cann't do it
by ourselves?

Any help is appreciated, thanks a lot.

• It seems that the pl file under the /usr/bin folder doesn't work. There's an error image named randomly such as e94c9a18c2cb40421bba8485dbb78af3.png would be created each time, but actually this image isn't showed up in the webpage, also not exist in the /images/timeline/ folder. Some ***.err like image would be created in the /images/timeline/ folder. When I checked each image folder, it seems that there's an EasyTimeline.txt.***** like file created in the /images/tmp folder each time when editing.

• Is the pl file form the pl240linuxi386.tar.gz package is just what mentioned in the EasyTimeline.pl file, an executable file?

Hello Jane,

I am aware of the EasyTimeline tool but have never used it.... you would need to
contact the author/maintainer of that tool for assistance.

-Steve
________________________________________
From: ploticus@yahoogroups.com [ploticus@yahoogroups.com] On Behalf Of
lemonlemon [sunflowerzf@...]
Sent: Sunday, November 09, 2008 5:43 AM
To: ploticus@yahoogroups.com
Subject: [ploticus] Problems for applying ploticus in our wiki site.

Hello,

Nice to join in this group.

I just found a very useful tool - EasyTimeline extension in Wikimedia.org, and
try to use it to draw in our mediawiki-powered site: http://www.cn0km.com/.

There's problems when installing. I just followed all instructions in the main
page, still no way to settle it.


  *   It seems that the pl file under the /usr/bin folder doesn't work. There's
an error image named randomly such as e94c9a18c2cb40421bba8485dbb78af3.png would
be created each time, but actually this image isn't showed up in the webpage,
also not exist in the /images/timeline/ folder. Some ***.err like image would be
created in the /images/timeline/ folder. When I checked each image folder, it
seems that there's an EasyTimeline.txt.***** like file created in the
/images/tmp folder each time when editing.

  *   Is the pl file form the pl240linuxi386.tar.gz package is just what
mentioned in the EasyTimeline.pl file, an executable file?

Any help is appreciated very much.

Great thanks.

Best Regards.

Jane.

-----Original Message-----
From: Steve Grubb [mailto:stevegrubb@...]
Sent: 12 November 2008 13:33
To: ploticus@...
Cc: ploticus@yahoogroups.com
Subject: Re: Minor issue re. date range spanning new-year

Thomas, thanks for this report.  I am aware of the EasyTimeline tool but don't use it, and am not familiar w/ exactly how it works.  The code shown in the raw view isn't a ploticus script.

I will check and see if the problem occurs with a plain vanilla ploticus script.

-Steve

---------- Forwarded message ----------


Hello,

EasyTimelinePlugin with ploticus installed and working.
When a date range spans new-year the tics can become mis-aligned.

> 1. First, make a fair effort towards learning the tool and solving the
problem yourself.

It looks like the tics are spaced based on a date-range days-count that is
one day off.

I first thought there might be something fishy in the overly complicated
juliandayno code in dates.c
but behaviour is unchanged after replacing it with known good code (provided
for completeness):

< static long
< _jul( y, m, d, offset )
< int y;                /* year */
< int m;                /* month */
< int d;                /* day */
< long offset;  /* calibrate for this julian day 0 */
< {
<     if (m++<3) { m += 12; --y; }
<     d += (int)(30.6*m)-63;
<     d += y*365+(int)(y/4)-(int)(y/100)+(int)(y/400);
<     return d-offset;
< }
<
< static int
< jul( y, m, d, julian )
< int y;                /* year */
< int m;                /* month */
< int d;                /* day */
< long *julian; /* julian date result */
< {
<     *julian = _jul( y, m, d, _jul( 1977, 1, 1, 0));
<     return 0;
< }



> 2. Include a stripped-down self-contained example that exhibits the
problem.

http://xenoc.demon.co.uk/twiki/bin/view/Interests/EasyTimelinePluginDev
http://xenoc.demon.co.uk/twiki/bin/view/Interests/EasyTimelinePluginDev?raw=
on

> 3. Use the -debug option .. and send the diagnostic results

N/A.

> 4. Indicate computing platform (Red Hat 7.2? Solaris 8? NT 4.0?)

CentOS 4.4

> 5. Indicate the exact pl command you used and describe exactly how you
invoked pl (command line? CGI?) If you use a config file please attach this
as well.

N/A.

> 6. Double check to be sure everything is correct. Did you attach the
correct script and data set? Did you describe the problem clearly and
succinctly?

Should be reproducible from the ploticus scripts embedded in the (raw view)
wiki page.

Kind Regards

--
Thomas

Hi,

I'm trying to build libploticus on a mac, but have run into some problems.  It
appears to
build correctly, but when I try to run pltestsuite/api_examp I get the following
error:

dyld: lazy symbol binding failed: Symbol not found: _PLGG_initstatic
   Referenced from: /usr/local/lib/libploticus.dylib
   Expected in: dynamic lookup


Any suggestions?  I built the library with gd, jpeg, png, and freetype support. 
The otool
output (showing linked libraries follows):

hahei:pltestsuite andy$ otool -L /usr/local/lib/libploticus.dylib
/usr/local/lib/libploticus.dylib:
	 /usr/local/lib/libploticus.dylib (compatibility version 0.0.0, current version
0.0.0)
	 /usr/local/lib/libgd.2.dylib (compatibility version 3.0.0, current version
3.0.0)
	 /usr/local/lib/libpng.3.dylib (compatibility version 3.0.0, current version
3.0.0)
	 /usr/lib/libz.1.dylib (compatibility version 1.0.0, current version 1.2.3)
	 /usr/local/libjpeg/lib/libjpeg.62.dylib (compatibility version 63.0.0, current
version
63.0.0)
	 /usr/local/lib/libfreetype.6.dylib (compatibility version 10.0.0, current
version
10.18.0)
	 /usr/lib/libgcc_s.1.dylib (compatibility version 1.0.0, current version 1.0.0)
	 /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version
111.1.1)


hahei:pltestsuite andy$ otool -L api_examp
api_examp:
	 /usr/local/lib/libploticus.dylib (compatibility version 0.0.0, current version
0.0.0)
	 /usr/local/lib/libpng.3.dylib (compatibility version 3.0.0, current version
3.0.0)
	 /usr/local/libjpeg/lib/libjpeg.62.dylib (compatibility version 63.0.0, current
version
63.0.0)
	 /usr/local/lib/libfreetype.6.dylib (compatibility version 10.0.0, current
version
10.18.0)
	 /usr/local/lib/libgd.2.dylib (compatibility version 3.0.0, current version
3.0.0)
	 /usr/lib/libz.1.dylib (compatibility version 1.0.0, current version 1.2.3)
	 /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version
111.1.1)
	 /usr/lib/libgcc_s.1.dylib (compatibility version 1.0.0, current version 1.0.0)

Please let me know if you have any idea what the problem may be.  Thanks!

--Andy

Stephen Grubb wrote:
>
>
> Hello Jane,
>
> I am aware of the EasyTimeline tool but have never used it.... you would
> need to contact the author/maintainer of that tool for assistance.
>
> -Steve

I posted some months ago a patch for ploticus to make it generate valid
xhtml (adding alt parameter to <area>), precisely targetted to the use
by EasyTimeline.
What's the status of it?

Hi, thanks for the reminder.  Your patch addressed one place where an area tag
was being generated in clickmap.c, but there was another place that it didn't
address.   I posted a revised version of clickmap.c to the web site that should
produce conformant code.  The change will be present in the next release.

-Steve
________________________________________
From: ploticus@yahoogroups.com [ploticus@yahoogroups.com] On Behalf Of
Platonides [platonides@...]
Sent: Wednesday, November 12, 2008 11:57 AM
To: ploticus@yahoogroups.com
Subject: Re: [ploticus] Problems for applying ploticus in our wiki site.

I posted some months ago a patch for ploticus to make it generate valid
xhtml (adding alt parameter to <area>), precisely targetted to the use
by EasyTimeline.
What's the status of it?

Stephen Grubb wrote:
>
> Hi, thanks for the reminder. Your patch addressed one place where an
> area tag was being generated in clickmap.c, but there was another place
> that it didn't address. I posted a revised version of clickmap.c to the
> web site that should produce conformant code. The change will be present
> in the next release.
>
> -Steve

Thank you very much.

-----Original Message-----
From: Steve Grubb [mailto:stevegrubb@...]
Sent: 12 November 2008 13:33
To: ploticus@...
Cc: ploticus@yahoogroups.com
Subject: Re: Minor issue re. date range spanning new-year

Thomas, thanks for this report.  I am aware of the EasyTimeline tool but don't use it, and am not familiar w/ exactly how it works.  The code shown in the raw view isn't a ploticus script.

I will check and see if the problem occurs with a plain vanilla ploticus script.

-Steve

---------- Forwarded message ----------


Hello,

EasyTimelinePlugin with ploticus installed and working.
When a date range spans new-year the tics can become mis-aligned.

> 1. First, make a fair effort towards learning the tool and solving the
problem yourself.

It looks like the tics are spaced based on a date-range days-count that is
one day off.

I first thought there might be something fishy in the overly complicated
juliandayno code in dates.c
but behaviour is unchanged after replacing it with known good code (provided
for completeness):

< static long
< _jul( y, m, d, offset )
< int y;                /* year */
< int m;                /* month */
< int d;                /* day */
< long offset;  /* calibrate for this julian day 0 */
< {
<     if (m++<3) { m += 12; --y; }
<     d += (int)(30.6*m)-63;
<     d += y*365+(int)(y/4)-(int)(y/100)+(int)(y/400);
<     return d-offset;
< }
<
< static int
< jul( y, m, d, julian )
< int y;                /* year */
< int m;                /* month */
< int d;                /* day */
< long *julian; /* julian date result */
< {
<     *julian = _jul( y, m, d, _jul( 1977, 1, 1, 0));
<     return 0;
< }



> 2. Include a stripped-down self-contained example that exhibits the
problem.

http://xenoc.demon.co.uk/twiki/bin/view/Interests/EasyTimelinePluginDev
http://xenoc.demon.co.uk/twiki/bin/view/Interests/EasyTimelinePluginDev?raw=on

> 3. Use the -debug option .. and send the diagnostic results

N/A.

> 4. Indicate computing platform (Red Hat 7.2? Solaris 8? NT 4.0?)

CentOS 4.4

> 5. Indicate the exact pl command you used and describe exactly how you
invoked pl (command line? CGI?) If you use a config file please attach this
as well.

N/A.

> 6. Double check to be sure everything is correct. Did you attach the
correct script and data set? Did you describe the problem clearly and
succinctly?

Should be reproducible from the ploticus scripts embedded in the (raw view)
wiki page.

Kind Regards

--
Thomas

The sample script file below demonstrates the problem.  Note that
bars A, B, and C have zero in column 4.  Yet, for me, the plot shows
a thin red line on the top of each bar.

Is it intended or my mistake?

The extraneous data would be covered up with outline:yes.  Also it
does not appear if I replace the "0" values with "-".

Duane W.

$ ploticus -v
usage: ploticus scriptfile [options] ...or...  ploticus -prefab prefabname
[options]
ploticus 2.40-Jan'08 (unix).  This build can produce: PS EPS SVG SVGZ PNG JPEG
WBMP FreeType2
Copyright 1998-2008 Steve Grubb, http://ploticus.sourceforge.net

#proc getdata
data:
    A 5 0 0
    B 0 4 0
    C 2 1 0

#proc categories
axis: x
datafield: 1

#proc areadef
rectangle: 1 1 5 5
yautorange: datafield=2,3,4 combomode=stack lowfix=0
xscaletype: categories

#proc bars
outline: no
lenfield: 2
locfield: 1
stackfields: *
color: brightgreen
legendlabel: A

#proc bars
outline: no
lenfield: 3
locfield: 1
stackfields: *
color: blue
legendlabel: B

#proc bars
outline: no
lenfield: 4
locfield: 1
stackfields: *
color: red
legendlabel: C

#proc settings
units:  cm
numbernotation: euro
#endproc

#loop

#proc page
landscape: yes
#endproc

#proc getdata
file:  testdata
nfields: 2
#endproc

#proc areadef
//areaname: whole
rectangle: 2 2 26 20
areacolor: yellow
title:  Meting
xautorange: datafield=1
yautorange: datafield=2 lowfix=0 nearest=10
#endproc

#proc xaxis
label:  sec
stubs:  inc
#endproc

#proc yaxis
label:  V
stubs:  inc
grid:  yes
#endproc

#proc lineplot
xfield:  1
yfield:  2
#endproc

#endloop

--- In ploticus@yahoogroups.com, "supra89ta" <speedtoys.racing@...> wrote:
>
> Im working with a # of datasets that I am now given on a weekly basis,
> where the datestamp is as follows for each datapoint:
>
> Sat Mar 01 11:13:16 PST 2008
> Sat Mar 01 11:14:16 PST 2008
> Sat Mar 01 11:15:16 PST 2008
> ...
>
>
> How do I get ploticus to use this dorky format as a date in a chron
chart?
>

Id like to bump this up..the list is getting way more attention now,
than it did in the spring.

Thanks.

Id like to bump this up..the list is getting way more attention now,
than it did in the spring.  Its still a problem..and I would love to
stay with ploticus.

Thanks.

--- In ploticus@yahoogroups.com, "supra89ta" <speedtoys.racing@...> wrote:
>
> Trying to use Ploticus to manage a large monthly dataset that I need
> to make charts of.
>
> Right now, I have to export these values from an application into
> individual files per file server, but I -can- export them as a single
> file, just that its very very long because the mass single export
> doesnt roll up each minute's sample into a single line, so the more
> file servers represented in the export file, the more massive the file
> gets as only a small percentage of the total output, has any actual
> values in it.
>
> SO..no probs, right?   I Seg fault on the OSX version I have here, on
> a windows version, a Linux version...all of them ive tried.
>
> Ideas?
>
>
>
> =================
> Running:
> ploticus 2.40-Jan'08 (unix).  This build can produce: PS EPS SVG X11
GIF
>
> OS:
> Up to date 10.5 OSX, MBP, 4G ram.
>
> Failing Command:
> plot f12.htm -maxrows 195610 -pagesize 10,6.5 -maxfields 20000000
> Note..that any value over 195609 segfaults.
>
> CSV File:
> -rw-r--r--  1 jmohler  jmohler  44027251 May  2 22:56 nfslat.csv
> [bipolar:new2/april-reports/6mo-reports] jmohler# wc -l *csv
>   869298 nfslat.csv
>
>
> Sample Data:  (Not including Row1, headers)
> 1,5109.294, , , , , , , , , , , , , , , , , ,
> 2, , , , , , , , , , ,5303.5, , , , , , , ,
> 3, , ,0, , , , , , , , , , , , , , , ,
> 4, , , , , , , , , , , , ,0, , , , , ,
> 5,4726.181, , , , , , , , , , , , , , , , , ,
> 6, , , , , , , , , , ,3110.075, , , , , , , ,
> 7, , ,0, , , , , , , , , , , , , , , ,
> 8, , , , , , , , , , , , ,0, , , , , ,
> 9,5822.402, , , , , , , , , , , , , , , , , ,
> 10, , , , , , , , , , ,5160.35, , , , , , , ,
>
>
>
> HTM FIle:
> #proc getdata
> file: nfslat.csv
> delimit: comma
>
> #proc areadef
>   areaname z0
>   title:  Filer12: Measured CPU and NFS Operational Latency
>   rectangle: 1 1 7.5 6
>   xautorange: datafield=1
>   yrange: 0 20000
>   xaxis.stubhide: yes
>   xaxis.tics: none
>   yaxis.stubs: inc 5000
>   yaxis.minortics: yes
>   yaxis.minorticinc: 1000
>   yaxis.label: NFS Latency
>   yaxis.labeldistance: .7
>
>  #proc drawcommands
>   commands:
>   color yellow
>   linetype 0 2 1
>   mov 1 4.75
>   lin 7.5 4.75
>   color red
>   mov 1 6
>   lin 7.5 6
>
> #proc annotate
>   verttext: no
>   location: 1. .8
>   textdetails: align=C
>   arrowhead: 60(s) 0.63(s)
> //  arrowtail: 60(s) 0.5(s)
>   text: Feb 10
>
> #proc annotate
>   verttext: no
>   location: 4.25 .8
>   textdetails: align=C
>   arrowhead: 4.25 1
>   text: Mar 10
>
> #proc annotate
>   verttext: no
>   location: 7.5 .8
>   textdetails: align=C
>   arrowhead: 7.5 1
>   text: Apr 10
>
>  #proc drawcommands
>   commands:
>   color black
>   linetype 0 1 1
>   mov 1 1
>   lin 7.5 1
>
> #proc curvefit
>   xfield: 1
>   yfield: 2
>   curvetype: movingavg
>   order: 1000
>   linedetails: color=blue width=0.5
>   legendlabel: NFS average (1Day)
>   maxinpoints: 5000000
>
> #proc curvefit
>   xfield: 1
>   yfield: 2
>   curvetype: movingavg
>   order: 15000
>   linedetails: color=green width=0.5
>   legendlabel: NFS average (30Days)
>   maxinpoints: 5000000
>
> #proc legend
>   location: max+1.1 max
>

Hello Duane,

see the proc bars "hidezerobars" attribute.   If you add a "hidezerobars: yes"
statement to the last proc bars invocation, this should make it behave as you
want.

Steve
________________________________________
From: ploticus@yahoogroups.com [ploticus@yahoogroups.com] On Behalf Of Duane
Wessels [9a68@...]
Sent: Monday, November 24, 2008 5:09 PM
To: ploticus@yahoogroups.com
Subject: [ploticus] Stacked bars with no outline show extraneous data when last
data column is zero

The sample script file below demonstrates the problem. Note that
bars A, B, and C have zero in column 4. Yet, for me, the plot shows
a thin red line on the top of each bar.

Is it intended or my mistake?

The extraneous data would be covered up with outline:yes. Also it
does not appear if I replace the "0" values with "-".

Duane W.

$ ploticus -v
usage: ploticus scriptfile [options] ...or... ploticus -prefab prefabname
[options]
ploticus 2.40-Jan'08 (unix). This build can produce: PS EPS SVG SVGZ PNG JPEG
WBMP FreeType2
Copyright 1998-2008 Steve Grubb, http://ploticus.sourceforge.net

#proc getdata
data:
A 5 0 0
B 0 4 0
C 2 1 0

#proc categories
axis: x
datafield: 1

#proc areadef
rectangle: 1 1 5 5
yautorange: datafield=2,3,4 combomode=stack lowfix=0
xscaletype: categories

#proc bars
outline: no
lenfield: 2
locfield: 1
stackfields: *
color: brightgreen
legendlabel: A

#proc bars
outline: no
lenfield: 3
locfield: 1
stackfields: *
color: blue
legendlabel: B

#proc bars
outline: no
lenfield: 4
locfield: 1
stackfields: *
color: red
legendlabel: C

Hello Henk,

I don't fully understand what you're doing, but did you try adding a

#proc page

Steve

________________________________________
From: ploticus@yahoogroups.com [ploticus@yahoogroups.com] On Behalf Of Steve
Grubb [stevegrubb1@...]
Sent: Friday, November 28, 2008 2:11 PM
To: ploticus@yahoogroups.com
Subject: [ploticus] Fwd: Redraw problem with x11 output in ploticus

Hello Steve,

You may skip this first part of my e-mail because it may make you blush a
little bit. But I must say I am impressed with Ploticus. I think of myself as
an old hand in computer graphics and supported GINO-F, GINO-GRAPH and the
like. These may not be known to you, but they where rather sophisticated
drawing subroutine packages (for  FORTRAN) since the beginning of the 1970s.
They still do exist. So I know a little bit of the problems you encounter and
the efforts you have put in Ploticus.

But now to the problem I encountered. I refer to the enclosed 'plotmeter'
script. You will see that it has an eternal loop. Every time I click
the  'More ...' button in the X11 window, the graph is refreshed. This is
done because the data in 'testdata' grows over time. About every half second
a new record is added by a multimeter coupled to the system via USB. The
result is a more or less 'real time' growing graph.

This works perfect but for one thing: part of the right side of the drawing
area is not cleared before the new plot is drawn. As long as the Xaxis stays
the same between plots this is not much of a problem, but you can see this
happen when, in a flash, part of the window turns white, with a yellow column
staying at the right. When the graph grows out of the boundary at the right
side, ploticus calculates a new Xaxis. Then the new axis is draw over the
partly still displayed old Xaxis. This can be seen in the enclosed png.
Somewhere between the old and the new number 800 is the boundary between the
cleared and the uncleared part. At th right side of this the numbers of
several 'old' Xaxis are still there to be seen.

Making a file testdata with:
    7    10
    6    20
and running 'pl plotmeter' will show the first plot. Then (leaving pl
runnning) add something to testdata:
    5    100
(the second number must be greater then the last value showed on the Xaxis to
force a rescaled Xaxis) and click More ..., will show what I mean.

This is nott a real showstopper, but I hope that this bugreport will help in
improving Ploticus. I am of course very willing to provide any information
you may ask for.

Some more:
Running pl on openSUSE 10.3 with KDE 3.5.
Installed ploticus from the openSUSE Build Service:
http://download.opensuse.org/repositories/home:/mge1512:/tools/openSUSE_10.3

--
regards,
Henk van Velden

Hello Jeff,

For both of the issues you wrote about, you could probably solve these problems
by processing the data before handing it to ploticus. Ploticus includes some
data manipulation capabilitiy but cannot do every imagingable operation,
especially using prefabs, which are meant for convenience in exploratory work.

Eg, If I was given the dorky date formats problem I would  write a shell script
that would send the data thru awk (or similar) to rearrange/concatenate the
fields to get them into a form that the chron prefab expects, or perhaps use a
ploticus script (which gives a lot more flexibility in this regard) instead of
using the chron prefab.

If I was encountering the problem with the large CSV file I would try to find a
way to reduce the size of the file or change it to tab-delimited, and try to
make progress that way.

Steve


________________________________________
From: ploticus@yahoogroups.com [ploticus@yahoogroups.com] On Behalf Of Jeff
Mohler [speedtoys.racing@...]
Sent: Friday, November 28, 2008 2:25 PM
To: ploticus@yahoogroups.com
Subject: [ploticus] Re: Ploticus Seg Faults on large CSV file

Id like to bump this up..the list is getting way more attention now,
than it did in the spring. Its still a problem..and I would love to
stay with ploticus.

Thanks.

--- In ploticus@yahoogroups.com<mailto:ploticus%40yahoogroups.com>, "supra89ta"
<speedtoys.racing@...> wrote:
>
> Trying to use Ploticus to manage a large monthly dataset that I need
> to make charts of.
>
> Right now, I have to export these values from an application into
> individual files per file server, but I -can- export them as a single
> file, just that its very very long because the mass single export
> doesnt roll up each minute's sample into a single line, so the more
> file servers represented in the export file, the more massive the file
> gets as only a small percentage of the total output, has any actual
> values in it.
>
> SO..no probs, right? I Seg fault on the OSX version I have here, on
> a windows version, a Linux version...all of them ive tried.
>
> Ideas?
>
>
>
> =================
> Running:
> ploticus 2.40-Jan'08 (unix). This build can produce: PS EPS SVG X11
GIF
>
> OS:
> Up to date 10.5 OSX, MBP, 4G ram.
>
> Failing Command:
> plot f12.htm -maxrows 195610 -pagesize 10,6.5 -maxfields 20000000
> Note..that any value over 195609 segfaults.
>
> CSV File:
> -rw-r--r-- 1 jmohler jmohler 44027251 May 2 22:56 nfslat.csv
> [bipolar:new2/april-reports/6mo-reports] jmohler# wc -l *csv
> 869298 nfslat.csv
>
>
> Sample Data: (Not including Row1, headers)
> 1,5109.294, , , , , , , , , , , , , , , , , ,
> 2, , , , , , , , , , ,5303.5, , , , , , , ,
> 3, , ,0, , , , , , , , , , , , , , , ,
> 4, , , , , , , , , , , , ,0, , , , , ,
> 5,4726.181, , , , , , , , , , , , , , , , , ,
> 6, , , , , , , , , , ,3110.075, , , , , , , ,
> 7, , ,0, , , , , , , , , , , , , , , ,
> 8, , , , , , , , , , , , ,0, , , , , ,
> 9,5822.402, , , , , , , , , , , , , , , , , ,
> 10, , , , , , , , , , ,5160.35, , , , , , , ,
>
>
>
> HTM FIle:
> #proc getdata
> file: nfslat.csv
> delimit: comma
>
> #proc areadef
> areaname z0
> title: Filer12: Measured CPU and NFS Operational Latency
> rectangle: 1 1 7.5 6
> xautorange: datafield=1
> yrange: 0 20000
> xaxis.stubhide: yes
> xaxis.tics: none
> yaxis.stubs: inc 5000
> yaxis.minortics: yes
> yaxis.minorticinc: 1000
> yaxis.label: NFS Latency
> yaxis.labeldistance: .7
>
> #proc drawcommands
> commands:
> color yellow
> linetype 0 2 1
> mov 1 4.75
> lin 7.5 4.75
> color red
> mov 1 6
> lin 7.5 6
>
> #proc annotate
> verttext: no
> location: 1. .8
> textdetails: align=C
> arrowhead: 60(s) 0.63(s)
> // arrowtail: 60(s) 0.5(s)
> text: Feb 10
>
> #proc annotate
> verttext: no
> location: 4.25 .8
> textdetails: align=C
> arrowhead: 4.25 1
> text: Mar 10
>
> #proc annotate
> verttext: no
> location: 7.5 .8
> textdetails: align=C
> arrowhead: 7.5 1
> text: Apr 10
>
> #proc drawcommands
> commands:
> color black
> linetype 0 1 1
> mov 1 1
> lin 7.5 1
>
> #proc curvefit
> xfield: 1
> yfield: 2
> curvetype: movingavg
> order: 1000
> linedetails: color=blue width=0.5
> legendlabel: NFS average (1Day)
> maxinpoints: 5000000
>
> #proc curvefit
> xfield: 1
> yfield: 2
> curvetype: movingavg
> order: 15000
> linedetails: color=green width=0.5
> legendlabel: NFS average (30Days)
> maxinpoints: 5000000
>
> #proc legend
> location: max+1.1 max
>

On Mon, 1 Dec 2008, Stephen Grubb said:

> Hello Duane,
>
> see the proc bars "hidezerobars" attribute.

doh!  sorry for the false alarm.

Hello Andreas,

there has never been any provision made in the ploticus code for handling the
jump to / from DST.

one suggestion would be to standardize your datetime values to  GMT

Sorry...

Steve

________________________________________
From: ploticus@yahoogroups.com [ploticus@yahoogroups.com] On Behalf Of Steve
Grubb [stevegrubb1@...]
Sent: Thursday, December 04, 2008 11:21 AM
To: ploticus@yahoogroups.com
Subject: [ploticus] How to handle daylight saving time

From: Andreas Müller

Date: Thu, Dec 4, 2008 at 11:02 AM
Subject: How to handle daylight saving time
To: stevegrubb@...<mailto:stevegrubb@...>


Steve,

how  do  I  have to format data with high time resolution so that
ploticus correctly shows the  time  discontinuity  occuring  e.g.
during the last weekend in october, when there certain timestamps
accur multiple times?  I found nothing about this problem on  the
ploticus website.

Mit herzlichem Gruss

                                        Andreas Müller

Hello Josh,

That's correct... maxrows needs to be set to accomodate the number of rows in
the data file.

While it's certainly possible there's a ploticus bug that makes an appearance
when a very large CSV file is read, it could be a problem with your data.  I
would suspect that perhaps there is something unusual on or near line 195,610 of
your data file.  Have you verified that your 800,000 row file is clean and
doesn't contain any malformed rows?

Does pl crash when you give it a 200,000 row CSV file?

The only way I can begin to look into a possible problem is to have a test case
and then try to reproduce the problem here.  That would mean you'd have to send
me the 800K row data set in question.... if you are sure it's clean and OK.  But
please check it first, especially around line 195,610

Steve

________________________________________
From: ploticus@yahoogroups.com [ploticus@yahoogroups.com] On Behalf Of Joshua
Simons [jsimons@...]
Sent: Wednesday, December 03, 2008 1:24 PM
To: ploticus@yahoogroups.com
Subject: Re: [ploticus] Re: Ploticus Seg Faults on large CSV file

Perhaps I'm not understanding, but the answer to this problem seems
to be included in the information supplied to document this issue.

Specifically, while the maximum number of data rows was increased to
195610, the input data file has over 800K rows in it. Doesn't maxrows
need to be at least as large as the number of rows in the data file?

Josh


On Dec 1, 2008, at 2:11 PM, Stephen Grubb wrote:


Hello Jeff,

For both of the issues you wrote about, you could probably solve these problems
by processing the data before handing it to ploticus. Ploticus includes some
data manipulation capabilitiy but cannot do every imagingable operation,
especially using prefabs, which are meant for convenience in exploratory work.

Eg, If I was given the dorky date formats problem I would write a shell script
that would send the data thru awk (or similar) to rearrange/concatenate the
fields to get them into a form that the chron prefab expects, or perhaps use a
ploticus script (which gives a lot more flexibility in this regard) instead of
using the chron prefab.

If I was encountering the problem with the large CSV file I would try to find a
way to reduce the size of the file or change it to tab-delimited, and try to
make progress that way.

Steve

_____________________ ___________________
From: ploticus@yahoogroups.com<mailto:ploticus%40yahoogroups.com>
[ploticus@yahoogroups.com<mailto:ploticus%40yahoogroups.com>] On Behalf Of Jeff
Mohler [speedtoys.racing@...<mailto:speedtoys.racing%40gmail.com>]
Sent: Friday, November 28, 2008 2:25 PM
To: ploticus@yahoogroups.com<mailto:ploticus%40yahoogroups.com>
Subject: [ploticus] Re: Ploticus Seg Faults on large CSV file

Id like to bump this up..the list is getting way more attention now,
than it did in the spring. Its still a problem..and I would love to
stay with ploticus.

Thanks.

--- In
ploticus@yahoogroups.com<mailto:ploticus%40yahoogroups.com><mailto:ploticus<mail
%20to:ploticus>%40yahoogroups.com>, "supra89ta" <speedtoys.racing@...> wrote:
>
> Trying to use Ploticus to manage a large monthly dataset that I need
> to make charts of.
>
> Right now, I have to export these values from an application into
> individual files per file server, but I -can- export them as a single
> file, just that its very very long because the mass single export
> doesnt roll up each minute's sample into a single line, so the more
> file servers represented in the export file, the more massive the file
> gets as only a small percentage of the total output, has any actual
> values in it.
>
> SO..no probs, right? I Seg fault on the OSX version I have here, on
> a windows version, a Linux version...all of them ive tried.
>
> Ideas?
>

Hello,

I'm using ploticus to plot some large set of data that contain IP addresses (as
numeric values)
and a value. IP addresses are on the x axis and the other values on the y axis.
My problem is
that, no matter how much I searched, I cannot find a way to print the x axis
stubs in a more
friendly way than a number in scientific notion (e.g. 2.5e+09).

It would much more convenient to print IP addresses in dotted format
(XXX.XXX.XXX.XXX ) or
even something like STUB / 2^24  (class A subnet).

thank you very much in advance,
vasilis

NOTICE: If received in error, please destroy and notify sender. Sender does not intend to waive confidentiality or privilege. Use of this email is prohibited when received in error.

Hello Wesley,

1. I think it can plot time values of format  hh:mm:ss.sss  or mm:ss.sss but
there may be issues with degenerate cases... it will probably work ok if you
have scattered ms values across at least several minutes.  It will probably not
work ok if you are trying to plot a 200ms time interval using time values.   It
may also depend on whether your data are elapsed time values,  clock time
values, or datetime values (where you might transtition from 23:59:59.999 to
0:00:00.00).  I would recommend downloading it and spending a little time to see
if you can get it to do what you want.

2. Ploticus can be used to create various types high-low bars and candlesticks
.. for example  http://ploticus.sourceforge.net/gallery/candlesticks.htm

Steve
________________________________________
From: ploticus@yahoogroups.com [ploticus@yahoogroups.com] On Behalf Of Bourne,
Wesley (IT) [Wesley.Bourne@...]
Sent: Monday, December 08, 2008 6:55 AM
To: ploticus@yahoogroups.com
Subject: [ploticus]

Hi,
I’m considering using Ploticus in an application and I thought I could ask you a
couple of quick questions as I couldn’t find the results in the read I gave your
documents.

1. Can Ploticus allow a time axis to be up to millisecond precision?
2. Can it supply finance bars/candle sticks
(http://t16web.lanl.gov/Kawano/gnuplot/intro/style-e.html#others)

Regards,
Wesley
________________________________

NOTICE: If received in error, please destroy and notify sender. Sender does not
intend to waive confidentiality or privilege. Use of this email is prohibited
when received in error.

Hello Vasilis, you should be able to use    xaxis.stubformat: %.0f

...to at least get them out of scientific notation.

If you are writing a ploticus script you could use proc getdata's "filter"
attribute to rewrite the ip addresses with dots.

Assuming the numeric IP addresses all have the same # of digits...

eg for data like this:
123234567789  44395  27
989764352341  22435  32

.......you could use something like the script below.  -Steve

#proc getdata
file: myfile
filter:
##set part1 = $substring(  @@1, 1, 3 )
##set part2 = $substring(  @@1, 4, 3 )
##set part3 = $substring(  @@1, 7, 3 )
##set part4 = $substring(  @@1, 10, 3 )
##set ip = @@part1 "." @@part2 "." @@part3 "." @@part4
##print @@ip  @@2   @@3

pf_fieldnames: ip count count2
showresults: yes


#proc categories
axis: x
datafield: 1

#proc areadef
yrange: 0 50000
xrange: categories
yaxis.stubs: inc
xaxis.stubs: usecategories

#proc bars
   locfield: ip
   lenfield: count
   color: red

________________________________________
From: ploticus@yahoogroups.com [ploticus@yahoogroups.com] On Behalf Of
bill_pappas3 [bill_pappas3@...]
Sent: Friday, December 05, 2008 6:11 PM
To: ploticus@yahoogroups.com
Subject: [ploticus] formating numeric axis stubs as IP addresses

Hello,

I'm using ploticus to plot some large set of data that contain IP addresses (as
numeric values)
and a value. IP addresses are on the x axis and the other values on the y axis.
My problem is
that, no matter how much I searched, I cannot find a way to print the x axis
stubs in a more
friendly way than a number in scientific notion (e.g. 2.5e+09).

It would much more convenient to print IP addresses in dotted format
(XXX.XXX.XXX.XXX ) or
even something like STUB / 2^24 (class A subnet).

thank you very much in advance,
vasilis

Hi,

I'm tabulating and plotting the number of events per hours of a day.
The input data looks like this:

   00:00:06
   00:00:11
   00:00:13
   00:00:13
   00:00:27
   00:00:31
   00:00:31
   ...
   23:59:28
   23:59:28
   23:59:31
   23:59:38
   23:59:38
   23:59:38
   23:59:38

And I'm calling ploticus like this:

   ploticus -eps -font /BasicCommercialLT-Light -tightcrop \
     -o submissions_ss2008_by_hour.eps -prefab chron \
     data=data_ss2008_by_hour  x=1 tab=hour unittype=time xinc="1 hour" \
     nearest=hour barwidth=0.15 stubfmt=HH xlbl=Hours ylbl=Submissions \
     ygrid=yes

Now I just noticed that the x axis is labeled from "00" to "00," whereas
I think it should be labeld from "00" to "24" since it is the end of the
same day, not the beginning of the next one.

Is there a way to achieve this?

Thanks and greetings


--
Michael Piotrowski, M.A.                               <mxp@...>
Public key at <http://www.dynalabs.de/mxp/pubkey.txt> (ID 0x1614A044)

If I set maxrows to more rows that the file has, it still segfaults.

Here is a sample of the data where it crashes in the csv file:


195607,,,,,,,,,,,,,,3012.764,,,,,
195608,,0,,,,,,,,,,,,,,,,,
195609,2857.373,,,,,,,,,,,,,,,,,,
195610,,,,0,,,,,,,,,,,,,,,
195611,,,,,,,,,,,,,12315.914,,,,,,



If I stop at 195609, im good, if I set maxrows to one more row..it seg faults.



On Thu, Dec 4, 2008 at 10:30 AM, Joshua Simons <jsimons@...> wrote:
> From what you say, the sole problem is likely that maxrows has not been set
> correctly. Failure to do so will cause the Ploticus read routine to scribble
> on
> other data structures, which could certainly cause the reported segfault.
> I am surprised, though, that Ploticus doesn't catch the fact that it has
> been
> asked to read more rows than it can accomodate based on maxrows
> setting.
> Josh
> On Dec 4, 2008, at 12:53 PM, Stephen Grubb wrote:
>
> Hello Josh,
>
> That's correct... maxrows needs to be set to accomodate the number of rows
> in the data file.
>
> While it's certainly possible there's a ploticus bug that makes an
> appearance when a very large CSV file is read, it could be a problem with
> your data. I would suspect that perhaps there is something unusual on or
> near line 195,610 of your data file. Have you verified that your 800,000 row
> file is clean and doesn't contain any malformed rows?
>
> Does pl crash when you give it a 200,000 row CSV file?
>
> The only way I can begin to look into a possible problem is to have a test
> case and then try to reproduce the problem here. That would mean you'd have
> to send me the 800K row data set in question.... if you are sure it's clean
> and OK. But please check it first, especially around line 195,610
>
> Steve
>
> ________________________________________
> From: ploticus@yahoogroups.com [ploticus@yahoogroups.com] On Behalf Of
> Joshua Simons [jsimons@...]
> Sent: Wednesday, December 03, 2008 1:24 PM
> To: ploticus@yahoogroups.com
> Subject: Re: [ploticus] Re: Ploticus Seg Faults on large CSV file
>
> Perhaps I'm not understanding, but the answer to this problem seems
> to be included in the information supplied to document this issue.
>
> Specifically, while the maximum number of data rows was increased to
> 195610, the input data file has over 800K rows in it. Doesn't maxrows
> need to be at least as large as the number of rows in the data file?
>
> Josh
>
> On Dec 1, 2008, at 2:11 PM, Stephen Grubb wrote:
>
> Hello Jeff,
>
> For both of the issues you wrote about, you could probably solve these
> problems by processing the data before handing it to ploticus. Ploticus
> includes some data manipulation capabilitiy but cannot do every imagingable
> operation, especially using prefabs, which are meant for convenience in
> exploratory work.
>
> Eg, If I was given the dorky date formats problem I would write a shell
> script that would send the data thru awk (or similar) to
> rearrange/concatenate the fields to get them into a form that the chron
> prefab expects, or perhaps use a ploticus script (which gives a lot more
> flexibility in this regard) instead of using the chron prefab.
>
> If I was encountering the problem with the large CSV file I would try to
> find a way to reduce the size of the file or change it to tab-delimited, and
> try to make progress that way.
>
> Steve
>
> _____________________ ___________________
> From: ploticus@yahoogroups.com<mailto:ploticus%40yahoogroups.com>
> [ploticus@yahoogroups.com<mailto:ploticus%40yahoogroups.com>] On Behalf Of
> Jeff Mohler
> [speedtoys.racing@...<mailto:speedtoys.racing%40gmail.com>]
> Sent: Friday, November 28, 2008 2:25 PM
> To: ploticus@yahoogroups.com<mailto:ploticus%40yahoogroups.com>
> Subject: [ploticus] Re: Ploticus Seg Faults on large CSV file
>
> Id like to bump this up..the list is getting way more attention now,
> than it did in the spring. Its still a problem..and I would love to
> stay with ploticus.
>
> Thanks.
>
> ---
> In
ploticus@yahoogroups.com<mailto:ploticus%40yahoogroups.com><mailto:ploticus<mail
> %20to:ploticus>%40yahoogroups.com>, "supra89ta" <speedtoys.racing@...>
> wrote:
>>
>> Trying to use Ploticus to manage a large monthly dataset that I need
>> to make charts of.
>>
>> Right now, I have to export these values from an application into
>> individual files per file server, but I -can- export them as a single
>> file, just that its very very long because the mass single export
>> doesnt roll up each minute's sample into a single line, so the more
>> file servers represented in the export file, the more massive the file
>> gets as only a small percentage of the total output, has any actual
>> values in it.
>>
>> SO..no probs, right? I Seg fault on the OSX version I have here, on
>> a windows version, a Linux version...all of them ive tried.
>>
>> Ideas?
>>
>
>