Roughly speaking, there are four kinds ROOT files in Minos:-
Rerooted file
e.g.: gm_far_le_noosc.root.
A "raw" Monte-Carlo file. This contains only Fortran-like objects. It
is read by a root job that has:-
jc.Input.Set("Format=reroot")
This can then be run through DetSim or RerootExodus to get something
that looks like raw data.
Raw data file:
e.g.: F000006101_0000.mdaq.root
Where F is for Far Detector (also C for CalDet).
The two numbers are the run and sub-run numbers, in
this case the first sub-run of run 6101.
This contains only the RawData blocks: raw detector data
from several streams e.g. DaqSnarl, DaqMonitor, LightInjection.
Reconstruction Candidate data file
The production jobs produce these with an extension .cand..root, e.g.
F00006101_0000.cand.R1.0.0a.root
This contains candidate objects, like CandTracks and CandShowers.
Standard ntuple file
This is a file with just a single TTree designed for making plots. It
is usually the ultimate output of a reconstruction analysis.
Currently there are actually 4 different ntuple files:-
- Extension .ntup..root (e.g. F00006101_0000.ntup.R1.0.0a.root)
Equivalent of eventsr.root This is an "old" style ntuple.
- Extension .ntps..root (e.g. F00006101_0000.ntps.R1.0.0a.root)
Equivalent of eventsr.sub.root (a slimmed down version of
eventsr.root.) This is an "old" style ntuple.
- Extension .sntp..root (e.g. F00006101_0000.sntp.R1.0.0a.root)
This is an "new" style ntuple.
- Extension .snts..root (e.g. F00006101_0000.snts.R1.0.0a.root)
This is an "new" style ntuple.
The plan is to phase out 1,2 in favor of 3,4. A difference between
1,2 and 3,4, is that while both sets of ntuple data can be viewed in
a basic root session using TBrowser, etc., data stored in 1,2 cannot
be read back into a subsequent minos job, while data in 3,4 can.
The above was plagiarised from an email exchange between
Nathaniel and Sue
|
