Hi. I've found, with the help of Bill Love, that the problem with fNdata=2 instead of 1 goes away for my primary vertex data (once per event), if I DON'T USE TClonesArray for the once per event information. The trade-off is that not using TClonesArray means the whole primary vertex (or event header) class is read in for each event. That's not a big deal when evaluating thousands of secondary vertices per event, whereupon most of the time is spent with the secondary vertices, but it is a big deal if I have a large collection of events with only a few vertices each. Especially if I start adding more and more info to the event header. My preference would be to fix the loophole in the TClonesArray case. -Gene Van Buren
This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:23 MET