Re: [ROOT] problem with tree->Draw()

From: Rene Brun (Rene.Brun@cern.ch)
Date: Fri Sep 01 2000 - 16:38:34 MEST


Hi HP,
When you use the option "goff", no graphics functions are called.
The fix I reported this morning was in the axis painting functions.
I cannot reproduce the problem with "goff".
Could you try to reproduce the problem with the file you sent to me
yesterday and give me the exact sequence of instructions to reproduce
the problem.

Rene Brun

HP Wei wrote:
> 
> Thanks for looking into the problem, Rene.
> Just to make sure that the fix you did completely settle this
> issue. Let me add more info.
> 
> >When looking at the file you sent to me yesterday, I see that your Tree
> >attributes have very large positive and negative values ( > 1e33).
>   Yes, there are a lot of -1.5092e33 in the data file.
>   In financial data, NaN is very common.
>   I use -1.5092e33 (-1/Planck const) to represent NaN.
> 
>   In the situation that I reported yesterday, (ntic==2) seems to be the one
>   causing the problem.
>   Using T->Scan("capt", "ntic==2"), I find that all the
>   entries for capt are NaN.  The Row numbers that T->Scan() generate
>   tells me that all these entries are scattered (in a systematic way, though)
>   around in the data file.
> 
>   -------------
>   By the way, is there a standard way to represent NaN in ROOT ??
>   For example, the package Matlab defines NaN as
>   double mxGetNaN()
> 
> >If one draws one of this attributes, the axis painting routine crashes
> >in the particular range of values that are in your file.
>   Does the option "goff" used in T->Draw() still trigger
>   any painting routines ???  I was using T->Draw()
>   with "goff" to extract
>   date into V1, V2 and V3.
>



This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:32 MET