Hi,
I put a prototype ntuple_branch.C into commis/phase1.
It makes "ntp" a TTree and split tdc and adc's into two branches.
This will solve problem of variable list string too long. We can add netron
counter to the ntuple with no difficulty at all. Since the
calibration of neutron counters is imminent, I think the solution is
urgently needed.
The current split into TDC branch and ADC branch is arbitrary. We can also
split it into LEFT and RIGHT all any way we deem appropriate.
The macro was run on 2347. if you open the TTree in a TBrowser or
TTreeViewer, you will see the structure. I tested, there is not problem
doing cross branch cuts. apply cut: "TDC:ttl0t>0 && TDC:ttl0t<4000" will
elliminate the pedestall in "ADC:atl0t".
Unfortunately, the Branch layer is not transparent. Old macros refering to
the elements simply by "ttl%dt" will need to change and use new convention
"TDC:ttl%dt". Therefore if or not to implement this split should be a
careful decision. Yet once the decision is made, I think ntuple_branch.C
demonstrated it is doable.
Since root processes TTress/Ntuples much faster than reading events from
coda files, I highly recommend using analysis codes on based on
ntuples/Trees.
I hope this will initiate more discussion on the matter. Possibly on the
analysis meeting tommorrow.
Chi
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:28 EST