Path: csiph.com!weretis.net!feeder6.news.weretis.net!feeder5.news.weretis.net!news.solani.org!.POSTED!not-for-mail
From: Karl Ratzsch <mail.kfr@gmx.net>
Newsgroups: comp.graphics.apps.gnuplot
Subject: Re: different behaviour gnuplot4 and 5 concerning "Implied independent variable y not found in fit function"
Date: Thu, 21 Mar 2019 22:02:21 +0100
Organization: solani.org
Lines: 45
Message-ID: <q70u4v$sjq$1@solani.org>
References: <e19bcee9-4a4b-42cf-b403-0ee850efd01c@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
X-Trace: solani.org 1553202143 29306 127.0.0.43 (21 Mar 2019 21:02:23 GMT)
X-Complaints-To: abuse@news.solani.org
NNTP-Posting-Date: Thu, 21 Mar 2019 21:02:23 +0000 (UTC)
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:60.0) Gecko/20100101 Thunderbird/60.6.0
In-Reply-To: <e19bcee9-4a4b-42cf-b403-0ee850efd01c@googlegroups.com>
X-User-ID: eJwFwQkBwEAIAzBL4yvYAXr4l7AkDIJNR8Dj4prOl6qFsnWSI98Wm2ibi2Ro7V77034Y+QE1/xJH
Cancel-Lock: sha1:sOwI5at0M9Q/haHPVuW1GyPf3EM=
Content-Language: de-DE-1901
Xref: csiph.com comp.graphics.apps.gnuplot:4169

Am 21.03.2019 um 16:02 schrieb hugocoolens@gmail.com:
> In a batch-file which worked fine for Gnuplot 4, I had the following lines:
> f(x) = a*10e-8*(exp(x/b)-1)
> err1(y)=0.01*0.033*y+2e-9
> err2(y)=0.01*0.0031*y+20e-9
> err3(y)=0.01*0.0034*y+200e-9
> err4(y)=0.01*0.0045*y+2e-6
> error(y)=(y<10e-6)? err1(y):(y<100e-6)? err2(y):(y<1e-3)? err3(y):err4(y)
> a=30e-3
> b=50e-3
> fit f(x) '1n4148_linstep_1na_10ma.dat' using ($1):($2):(error($2)) via a,b
> 
> This worked fine in Gnuplot 4, however when I try this in Gnuplot 5, I get the following error message:
> 	> Implied independent variable y not found in fit function.
> 	> Assuming version 4 syntax with zerror in column 3 but no zerror keyword.
> 

It's only a warning, not an error. Doesn't the fit converge?

I tried this:

   set xr [0:1e-3]
   fit f(x) '+' using ($1):(f($1)):(error(f($1))) via a,b

where gp of course can't compute errors from the perfect data, but
otherwise works fine.

To get rid of the warning, you specify to which variable the
additional error column refers:

      fit f(x) '1n4148_linstep_1na_10ma.dat' using
($1):($2):(error($2)) yerror via a,b

(you can also give "zerror" or "error z", that's all the same if you
only have one dependent and one independent variable.)

I understand the warning to be just one warning, not two: You gave
three columns, but the fit function only has one variable. That
makes gnuplot assume (correctly,) that you don't have two
independent variables, but one, plus an additional column with
y(==z)-errors.


   Best, Karl