Why is there a difference between ecdf and cdfcalc plots?

Hello,

I tried to plot a complentary cumulative distribution function based on some empircial data set using the ecdf function. Then, I found another possible solution to that specific problem by using the cdfcalc function that also seems to be the basic of the cdfplot function. The reason why I did not use the cdfplot function directly, is the fact that I was looking for a 1-cdf plot as mentioned before. Now that I have implemented both solutions, I get slightly different results as shown in the graph (one is shifted to the right). Since the output of the cdfcalc function returns two vectors of different size, the ycdf vector has to be cropped (according to a question asked in MATLAB Answers, see https://ch.mathworks.com/matlabcentral/answers/8368-plot-complementary-cumulative-distribution). Why is there a difference between both approaches and which one should I use?

%empircial data stored in vector hist
load('hist.mat','hist'); %attached hist.mat-file
hold on
%First approach using cdfcalc
[ycdf,xcdf] = cdfcalc(hist);
xccdf = xcdf;
yccdf = 1-ycdf(1:end-1);
stairs(xccdf,yccdf);
%Second approach using ecdf 
[f,x]=ecdf(hist);
fc=1-f;
stairs(x,fc);
hold off