Re: [ROOT] How to check if the histogram is empty?
From: Mateusz Ploskon (M.Ploskon@gsi.de)
Date: Fri Apr 25 2003 - 03:31:57 MEST
Hello Rene,
Thank You for the effort.
I'll deal with the problem the hard way... I'll compare the under and
over flow bins with the GetEntries() return - maybe in the future one
could add a function which returning = NUnderFlows+NOverFlows-NEntries
of a histogram(?) - if that's zero then histogram is "empty"...
(PS I fill the histo using Fill as in the example I've sent.)
Best regards,
Mateusz
attached mail follows:
Hi Mateusz,
In the current CVS version, I have modified the SetBinContent functions
to also increment the number of entries.
This should also fix the problem when drawing 2-d histograms
filled via SetBinContent.
Rene Brun
On
17 Apr 2003, Mateusz Ploskon wrote:
> Dear Rene,
>
> Thak You for Your prompt answer.
> I think my question was a little bit confusing...
> Let me clearify what I ment by an "empty" histogram...
> The GetEntries() returns also under and over flows while the "content"
> (i.e. data which might tell You something else than if the
> entry went outside the histogram limits...:)) - that is not interesting
> for me... I thought there might be a smart to check if thereĀ“s something
> inside the histo bounds.
>
> Please consider following:
> 1)
> {
> TH1F a("a","a",10,0,9);
> a.Fill(5, -1);
> a.Fill(6, 1);
> a.GetEntries(); // = 2
> a.GetSumOfWeights(); // = 0
> }
>
> 2)
> {
> TH1F a("a","a",10,0,9);
> a.Fill(-10000);
> a.Fill(100000);
> a.GetEntries(); // = 2
> a.GetSumOfWeights(); // = 0
> }
>
> ... but the pictures are not comparable... :(
>
> Of course one could have a look into Mean and RMS but, in strange cases
> I think even this might not help...
> There's also possibility to check the underflow && over bins but for
> histo dim > 1 this makes some fuss... while I iterate on the TList when
> checking... (fuss = I can call GetNbinsX(), ..Y, ..Z to get the numbers
> and check the under and overflow bins and compare with GetEntries and so
> on...)
>
>
> Best regards,
> Mateusz
>
This archive was generated by hypermail 2b29
: Thu Jan 01 2004 - 17:50:11 MET