Using loglog to plot results in a y axis that is not log scaled. But matlab seems to think it is

Hi, I've been trying to plot this data on a double log plot. However matlab seems to only produce a plot with the y axis on a linear scale instead of a log scale. And despite useing set(gca,'YScale','log') and all combinations of hold being on and off, nothing seems to work

I have tried both

frictiondata = [62118,11.037;59128,10.987;55209,10.919;49285,10.805;42407,10.655;35144,10.467;27427,10.219;17660,9.7791;10705,9.2785;8016,8.9892]; 

loglog(frictiondata(:,1),frictiondata(:,2),'ok')
set(gca,'YScale','log','XScale','log')
grid on

and

frictiondata = [62118,11.037;59128,10.987;55209,10.919;49285,10.805;42407,10.655;35144,10.467;27427,10.219;17660,9.7791;10705,9.2785;8016,8.9892]; 

hold on
loglog(frictiondata(:,1),frictiondata(:,2),'ok')
set(gca,'YScale','log','XScale','log')
grid on

But both versions of the plot give the same result shown in the figure below

Additionally, matlab seems to think that the y axis is on a log scale like the code would dictate. However, the y axis still appears to be on a linear scale on the genereated plot (again shown in the figure below).

I would also like to emphasize that a semilog plot would not be sufficient for my use.

Thank you very much for your time and effort!

[per isakson]

I think YScale actually is log and that the small range of your y-data makes it hard to observe it.

%%
frictiondata = [1.5,1.5; 10,10; 100,100; 0.7e3,0.7e3 ]; 
figure
loglog(frictiondata(:,1),frictiondata(:,2),'ok')
grid on

outputs

Answer 1

[Walter Roberson]

Your data is very nearly effectively linear in log x. sum((y-log(x)).^2) is 5.4E-7 -- a pretty low residual. You can do slightly better with 0.999927260504294 * log(x) - 0.000652400480442163 for a residue of 4.1E-7

Over that range of y values (8.9892 to 11.037), there is not much visual difference between log and linear. If you compare the two plots side by side you can see a bit of a difference in spacing of the ticks.

Answer 2

[dpb]

Set

ylim([5 15])

to have sufficient part of the y-axis showing to be able to see the difference in the log scale vis a vis linear.

Just have too small a data range for it to be obvious with the limits you've got.

That's also 5K-80K on x axis.

Even here the difference isn't terribly obvious in the data range, but between the bottom and top of the graph range it's clear the y scale is logarithmic.

[Gregory Gassen]

Great, Thank you, I ended up reallizing I was plotting the wrong data because of this. Again thanks!