[Halld-offline] DSelector: Generating Flat TTree

[Paul Mattione]

FYI. I just fixed a bug in how it was handling missing and decaying particles for the flat tree, so you may want to update your code again. 

 - Paul

----- Original Message -----
From: "Paul Mattione" <pmatt at jlab.org>
To: "GlueX Software Help" <gluex-software at googlegroups.com>, "GlueX Offline Software Email List" <halld-offline at jlab.org>
Sent: Friday, March 3, 2017 3:56:23 PM
Subject: Re: DSelector: Generating Flat TTree

Something I forgot to mention: For this to work, you also have to update your sim-recon, because I added a new function to the particleType.h file. 

 - Paul

----- Original Message -----
From: "Paul Mattione" <pmatt at jlab.org>
To: "GlueX Software Help" <gluex-software at googlegroups.com>, "GlueX Offline Software Email List" <halld-offline at jlab.org>
Sent: Friday, March 3, 2017 3:03:15 PM
Subject: DSelector: Generating Flat TTree

When running your DSelector, you can now optionally generate a flat version of the TTree (one entry = one combo, rather than one entry = one event).  Once you update your code, you can generate a new selector to see example code on how to do this. 

The intent is that this format can be used to easily slice-and-dice the data to see what happens with various cuts. However, since it's very difficult to handle combo-double-counting correctly in this format (and it's impossible to do so with PROOF), it's recommended that your final analysis still be done with the DSelector (just put your cuts in there).  

Note that I've only tested this for a few channels, so if you run into problems with it please let me know.  

 - Paul