Better way to autoscale x axis a histogram

I am trying to "autoscale" the x-axis of a histogram on an axes component. I think I have done it using for loops but was wondering if there was a more elegant way to do it.

I first get the histogram data from the axes.

axes(handles.axes2);  
barObj= findobj(gca, 'type', 'bar');
x=get(barObj,'XData')';
y=get(barObj,'YData')';

Determine max x & max y

mxy=max(y);
mxx=max(x);

Find index of max Y value so can get its "x" value. Also define a level value, below which I will declare the "bounds" of the histogram to then pass to xlim.

idx = find(y==mxy);
level=0.01*mxy

Get upper limit of X axis: The idea is to count x from the x location where Y is max and count outwards until 3 consecutive Y values less than level are obtained.

%Get upper limit
      for i=idx:max(x)-3
       i=i+1;
       y1=y(i);
          y2=y(i+1);
          y3=y(i+2);
              if(y1&y2&y3<level)
               indexH=i;
                 break
              end    
      end

Likewise for the lower X value bound:

%Get lower limit
    indexL=0;    %set incase not found
    for i=0:idx
        i=i+1;
        y1=y(i);
        y2=y(i+1);
        y3=y(i+2);
            if(y1&y2&y3>level)
               indexL=i;
            break
            end    
    end

Now rescale the x -axis

   xlim([indexL indexH])

Here is the original histogram

and what I am trying to achieve (more elegantly)

Answer 1

[Kirby Fears]

After plotting:

axis tight
axis 'auto y'

axis tight makes both x and y axes fit the min/max values of the axis. Then 'auto y' reverts the y axis scale back to the standard setting.

[Jason]

Hi, thanks for your suggestion. Unfortunately I have an outlier at a value of x=340 (see the top graph which is autoscaled in x). So your suggestion just duplicates the top image.

[Kirby Fears]

Have you considered excluding outliers before plotting? It makes sense to apply your "selection" logic to the data itself instead of the axis boundaries:

"Get upper limit of X axis: The idea is to count x from the x location where Y is max and count outwards until 3 consecutive Y values less than level are obtained."