Re: MyHist_2D ->Draw() :: Intersting

From: Marc Escalier <escalier_at_lal.in2p3.fr>
Date: Mon, 9 Aug 2010 22:21:18 +0200


Hi,

the fact that it works with TBrowser is not a proof : TBrowser could be a very evoluated class that have protection against some strange values and so on.

ok, so you say that TH1 with fabs(dx) works and 2D doesn't work

> I also draw 1D histogram of "dx" which is fine.
> myhist_1D->Fill(fabs(dx));

but i guess that if you make some prints, you could see something strange appear, maybe NotANumber, or something like this (maybe some absence of protection from sqrt(negative number) and so on

if you do that for the same size of sample and that you see anomalous values, then i don't know, and maybe experts of root will reply.

But if had your problems, i would have proceeded in 2 steps : -see if the problem occurs for some typical events (isolate events for which i see the strange behaviour)
-print the values to see if what i think enter the histogram is really what i expect

regards



On Mon, 9 Aug 2010, asaddiqu_at_ualberta.ca wrote:

> I am not on lxplus server. It's cdf stuff. "dt" and "dx" are well defined in
> cdf coding, like
>
> float dt = fCosmicBlock->DeltaTof();
> float dx = d0Corr(track1);
>
> But please don't forget that i can see these histograms in "TBrowser" . My
> only problem is "Draw()".
> I also draw 1D histogram of "dx" which is fine.
> myhist_1D->Fill(fabs(dx));
>
> If there is some problem in filling then i shouldn't be able to see in
> "TBrowser".
>
> Cheers, Asif
>
>
>
>
>
>
>
>
>
>
>
> Quoting "Marc Escalier" <escalier_at_lal.in2p3.fr>:
>
>> Hi,
>>
>> the behaviour that you observe is strange.
>>
>> ==> Maybe you should put some printing in your code to see what dt and
>> fabs(dx) are.
>>
>> maybe that fabs(dx) give you back a strange stuff : so printing it could
>> help for the investigation
>>
>> i assume that dt and dx are not pointers and not vectors.
>>
>> and maybe provide a minimal code to check (if feasible)
>>
>> and provide some copy/paste of the line that define :
>> myhist_2D
>> myhist_2D_abs
>> dt
>> dx
>>
>> (or also, put your code in a public directory of lxplus (i don't find it in
>> ~asaddiqu at lxplus)
>>
>> regards
>>
>> ====
>> On Mon, 9 Aug 2010, asaddiqu_at_ualberta.ca wrote:
>>
>>> Thank you Stephan & Marc for your reply.
>>>
>>> I am pretty sure that there is no sub directory problem. I just found an
>>> interesting thing which might explain my problem.
>>> Let say i have two parameters time "dt"(which is always positive) &
>>> distance from either side of a vertex"dx"(could be positive or negative)to
>>> fill a 2D histogram. In first case i take absolute value of "dx" and in
>>> second case i don't,
>>>
>>> Case1:
>>> myhist_2D_abs->Fill(dt,fabs(dx));
>>>
>>> Case2 :
>>> myhist_2D->Fill(dt,dx);
>>>
>>> So, "Draw()" command works for second case but does not work for first
>>> case(for absolute value) and gives an error message that i already
>>> mentioned.
>>>
>>>
>>> Cheers, Asif
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> Quoting asaddiqu_at_ualberta.ca:
>>>
>>>>
>>>> Hi,
>>>>
>>>> I can view my 2D histogram by "TBrowser" but whenever i try to see it by
>>>> "Draw()".
>>>> It shows an error message which is following.
>>>>
>>>> ============================================
>>>> root[1] myhist_2D->Draw()
>>>> Error: Symbol myhist_2D is not defined in current scope FILE:(tmpfile)
>>>> LINE:1
>>>> Error: Failed to evaluate myhist_2D->Draw()Possible candidates are...
>>>> filename line:size busy function type and name
>>>> =============================================
>>>>
>>>> Unfortunately, i am using an older version of root{4.00/08 December 2004
>>>> } but there should be no problem of version with this basic thing.
>>>>
>>>> Cheers, Asif
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>
>>
>
Received on Mon Aug 09 2010 - 22:21:33 CEST

This archive was generated by hypermail 2.2.0 : Tue Aug 10 2010 - 05:50:02 CEST