According to the documentation for TGraphErrors, there is a constructor that reads input text files. The documentation, http://root.cern.ch/root/html/TGraphErrors.html#TGraphErrors:TGraphErrors@8, states that a format of “%lg %lg %lg” will read the first three columns a as X, Y, and EY. When running the following script, errors are read into EX. As shown in the plot.
TGraphErrors* g= new TGraphErrors("test.csv", "%lg %lg %lg", ",");
g->Draw("AP");
Is this a mistake in documentation of implementation.
Exactly, the documentation does not make it clear that the functionality of this method changes when a delimiter is specified. Is this a documentation or implementation mistake?
The code is totally different. You can have a look if you want. I am not sure why the author did not implement it that way, but looking at the code it looks more complex.
Any way a global edit or a small “sed” command on your csv file to replace ; by blanck will do the job.
If you come with a working implementation we will be please to include it.