Hi Rene,
thanks a lot you will implement this! Here is a small macro that creates a
tree with a gaussian variable "a" and a bitfield "bits" which is a
TBits(64). The selected bit depends on the value of "a" from -2 to 2. E.g:
btest.Draw("a","bits.TestBitNumber(0)")
will give you all entries <2.
Cheers,
Birger
-------------------------------------------------------------
void makeBits()
{
TRandom r;
gROOT->Reset();
Float_t a;
TFile *foutP = new TFile("bittest.root","RECREATE");
TTree *outTreeP = new TTree("btest","Bits");
outTreeP->Branch("a",&a,"F");
TBits *myBits=new TBits(64);
outTreeP->Branch("bits","TBits",&myBits);
for(int i=0; i< 10000; i++){
float x;
x=gRandom.Gaus(0,1);
a=x;
x=x*16+32;
int j=(int) x;
if(j<0) j=0;
if(j>63) j=63;
myBits.ResetAllBits();
myBits.SetBitNumber(j);
outTreeP->Fill();
};
foutP->Write();
foutP->Close();
}
----------------------------------------------------------------
On Thu, 31 Oct 2002, Rene Brun wrote:
> Hi Birger,
>
> On Thu, 31 Oct 2002, Birger Koblitz wrote:
>
> > Hi Rene,
> >
> > yes, this is what we would like to have. Sorry for my poor description.
> > Just two additional clarifications:
> > - The operator [] should also work in selections:
> > mytree.Draw("somethingElse","myEventBits[10]");
> >
>
> Yes, this was implicit in my previous mail. In the selection, mask
> operations should also be supported, eg
> mytree.Draw("somethingElse","myEventBits & 0x3ff");
>
> > - Not only calling mytree.Draw("myEventBits"); should produce this
> > histogram, but also clicking on the variable in the TBrowser.
>
> Yes, it is the same logic. TBrowser calls TTree::Draw.
>
> OK, we will implement this logic. Could you send me a small TTree file
> (not HAT file) that we could use to test the new feature?
>
> Rene Brun
>
>
> >
> > Cheers,
> > Birger
> >
> >
> > On Thu, 31 Oct 2002, Rene Brun wrote:
> >
> > > Hi Birger,
> > >
> > > Let see if I understand correctly your request.
> > >
> > > - you are requesting an operator [] to get the ith bit in a TBits
> > > instead of calling the function TestBitNumber().
> > > That is quite easy to implement. I propose in this case
> > > a read only operator [].
> > >
> > > - you want to do mytree.Draw("myEventBits");
> > > This will automatically loop on all your events and filling
> > > an hitogram with as many channels as there are bits in TBits
> > > and where the bin count is incremented by 1 or the specified weight
> > > for each event where the corresponding bit is set.
> > > This requires a bit more work, but I agree would be a nice addition.
> > >
> > > Rene
> > >
> > > On Thu, 31 Oct 2002, Birger Koblitz wrote:
> > >
> > > > Hi Rene,
> > > >
> > > > I think Ursula Berthon already talked to you about this at the
> > > > Root-Meeting:
> > > >
> > > > Within H1 we would like to be able to put bit-arrays into our HAT tree
> > > > which should be accessible using only a browser and which is used for
> > > > event selection. Currently TBits allows to use the member function
> > > > TestBitNumber() which is quite inconvenient if it comes to a selection of
> > > > a large number of status bits e.g. to check if the quality of an event is
> > > > good. This could easily go over the maximum selection length of a tree.
> > > > Therefore it would be nice if one could simply use bits[i] to check the
> > > > ith bit in TBits bits. In addition it would be nice if one could view a
> > > > bit-array in a TBrowser. It should show a histogram over the bits and the
> > > > number of entries per bit. Currently fAllBits shows the entries (in 2^i
> > > > form) for the lowest byte. I think this is not really useful.
> > > >
> > > > In the moment we are using byte-arrays which at least allow to use the []
> > > > operator and allow us to inspect one bit with the TBrowser. H1 would be
> > > > very intersted to move to TBits because currently the bit-arrays make up
> > > > most of our HAT ntuple, we would be even willing to move to a newer
> > > > root-version ;-)
> > > >
> > > > Cheers,
> > > > Birger
> > > >
> > > > ------------------------------------------------------------
> > > > Birger Koblitz koblitz@mail.desy.de
> > > > MPI für Physik
> > > > c/o DESY-FH1 Tel: +49 40 8998-2431
> > > > Notkestr. 85 Fax: +49 40 8998-3094
> > > > D-22603 HAMBURG
> > > >
> > >
> > >
> >
> > ------------------------------------------------------------
> > Birger Koblitz koblitz@mail.desy.de
> > MPI für Physik
> > c/o DESY-FH1 Tel: +49 40 8998-2431
> > Notkestr. 85 Fax: +49 40 8998-3094
> > D-22603 HAMBURG
> >
>
>
------------------------------------------------------------
Birger Koblitz koblitz@mail.desy.de
MPI für Physik
c/o DESY-FH1 Tel: +49 40 8998-2431
Notkestr. 85 Fax: +49 40 8998-3094
D-22603 HAMBURG
This archive was generated by hypermail 2b29 : Sat Jan 04 2003 - 23:51:16 MET