gnuplot-info

[Gnuplot-info] Broken histograms when using smooth frequency on data file with discontinuities

[David L. <lin...@ie...>]

Hello,

I would like to plot the distribution of a variable and I follow the
directions given in the "Gnuplot in Action" book:

I define a binning function:

bin(x,s) = s*int(x/s)

And then I use the plot command:

plot './data.dat' u (bin($2, 0.2)):(1) smooth frequency

However, the 'data.dat' file has a number of discontinuities in it
(i.e. blank lines). This causes the output to show several sets of
histograms, one for each continuous set of data.

Is there any workaround for this? I would like the histogram to show
the distribution in the whole file, and to ignore discontinuities.

Thanks in advance,

-- 
::::::::::::::::::::::::::::::::::::::::::::
David Lindelöf, Ph.D.
+41 (0)79 415 66 41 or skype:david.lindelof
Better Software for Tomorrow's Cities:
  http://computersandbuildings.com
Follow me on Twitter:
  http://twitter.com/dlindelof

Re: [Gnuplot-info] Broken histograms when using smooth frequency on data file with discontinuities

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

David Lindelöf <lindelof <at> ieee.org> writes:

> I would like to plot the distribution of a variable and I follow the
> directions given in the "Gnuplot in Action" book:
> 
> I define a binning function:
> 
> bin(x,s) = s*int(x/s)
> 
> And then I use the plot command:
> 
> plot './data.dat' u (bin($2, 0.2)):(1) smooth frequency
> 
> However, the 'data.dat' file has a number of discontinuities in it
> (i.e. blank lines). This causes the output to show several sets of
> histograms, one for each continuous set of data.
> 
> Is there any workaround for this? I would like the histogram to show
> the distribution in the whole file, and to ignore discontinuities.

maybe use an external program to omit the blank lines:

plot '< awk "{if (NF) print}" ./data.dat' u (bin($2, 0.2)):(1) smooth frequency

"awk" should be installed on every unix system, for windows systems google
for 'awk windows'.

Re: [Gnuplot-info] Broken histograms when using smooth frequency on data file with discontinuities

[walter h. <wh...@bf...>]

Am 15.11.2011 17:08, schrieb Thomas Sefzick:
> David Lindelöf <lindelof <at> ieee.org> writes:
> 
>> I would like to plot the distribution of a variable and I follow the
>> directions given in the "Gnuplot in Action" book:
>>
>> I define a binning function:
>>
>> bin(x,s) = s*int(x/s)
>>
>> And then I use the plot command:
>>
>> plot './data.dat' u (bin($2, 0.2)):(1) smooth frequency
>>
>> However, the 'data.dat' file has a number of discontinuities in it
>> (i.e. blank lines). This causes the output to show several sets of
>> histograms, one for each continuous set of data.
>>
>> Is there any workaround for this? I would like the histogram to show
>> the distribution in the whole file, and to ignore discontinuities.
> 
> maybe use an external program to omit the blank lines:
> 
> plot '< awk "{if (NF) print}" ./data.dat' u (bin($2, 0.2)):(1) smooth frequency
> 
> "awk" should be installed on every unix system, for windows systems google
> for 'awk windows'.
> 
> 

 plot '< grep -v ^$ ./data.dat'  #seems more easy

just my 2 cents,
re,
 wh

Re: [Gnuplot-info] Broken histograms when using smooth frequency on data file with discontinuities

[David L. <lin...@ie...>]

I agree that would work. However, my data file happens to also have
several data sets in it (i.e. blocks of data separated by two blank
lines), and I would like to retain that information.

Personally I don't quite understand why plotting with boxes should
yield discontinuities when crossing a gap in the data. Isn't this
kinda lika bug of sorts?

On Mon, Nov 14, 2011 at 10:43 PM, Hans-Bernhard Bröker
<HBB...@t-...> wrote:
> On 14.11.2011 17:23, David Lindelöf wrote:
>
> [...]
>
>> plot './data.dat' u (bin($2, 0.2)):(1) smooth frequency
>>
>> However, the 'data.dat' file has a number of discontinuities in it
>> (i.e. blank lines). This causes the output to show several sets of
>> histograms, one for each continuous set of data.
>>
>> Is there any workaround for this? I would like the histogram to show
>> the distribution in the whole file, and to ignore discontinuities.
>
> Then you'll have to get rid of those discontinuities.  You can do that
> statically or on-the-fly, as a preprocessing step before sending the data to
> gnuplot.  Something like the following might work:
>
>        plot "< grep -v '^$' data.dat" u (bin($2,0.2)):(1.0) sm freq
>
>



-- 
::::::::::::::::::::::::::::::::::::::::::::
David Lindelöf, Ph.D.
+41 (0)79 415 66 41 or skype:david.lindelof
Better Software for Tomorrow's Cities:
  http://computersandbuildings.com
Follow me on Twitter:
  http://twitter.com/dlindelof