Re: THashTable of SuperBranch?

Rene Brun (Rene.Brun@cern.ch)
Mon, 15 Mar 1999 08:27:47 +0000

Tioukov Valeri wrote:
>
> Hi rooters,
>
> How to solve in the best way the following problem (I think very common):
>
> I have big bank of events ~30000 and I need the fast access to them by
> the event identifier.
>
> Now my events are stored in the Tree and it is very convinient for
> analysis. But then I need to get event by it's identifier it is not
> straightforward. One branch (leaf) of the tree is the eventID. So if I
> need to get the event NNNN the obvious way is to loop this branch by all entries
> upto eventID==NNNN. Due to split algorithme it is not slow, but
> for the big number of events the time penalty is significant especially
> if I need to do this operation 3000 times.
>
> To speedup the process I created for this tree the hashtable keeping the
> special objects with 2 numbers : eventID and Entry. Using this hashtable
> I can fastly find event Entry by the eventID and then get the
> event from Tree by the Entry.
> It works. But this solution seems to me not very elegant. May be in Root
> already exists something more automatic?
>
> Regards
> Valeri

Hi Valeri,
The THashTable is a good solution. However, it requires a table with
2 integers per entry. Another solution (may be faster) and more
economical
could be to build an array with the list of your eventIDs and use
the TMath::BinarySearch functions recently introduced in TMath.

Rene Brun