Unless you reference ALL variables in a query, it is always better
to store one variable per branch.
-the file size will be smaller because the compression algorithm
will work better on homegeneous variables
-you minimize the amount of I/O if your query references only
a subset of all the variables.
Note that we will come soon with a new class TTreeSQL that will allow you to query directly a SQL table as if it was a ROOT Tree.
Rene Brun
On Sat,
26 Mar 2005, Ivan Koji Koga wrote:
> Hello all,
> I have a database in mysql and i want to store it in ROOT TTree.
> I declared a ttree for one table and then i put all columns in a branch
> like:
>
> tree->Branch(“RoundTrip”, &roundtrip,
> “testid/I:microseconds:sourceid:destid:datetime/C:delay/F:jitter:loss”);
>
> What's the difference in performance if i do it with many branches like:
> tree->Branch("testid", &testid, "testid/I");
> tree->Branch("microseconds", µseconds, "microseconds/I");
> and so on...
> I saw in the userguide that if you have some related variables you have
> to put it together in the same branch. I think that my data is related
> because they are all in the same table and have values linked.
>
>
> Thank's in advance,
> Ivan K. Koga
>
Received on Sun Mar 27 2005 - 10:49:05 MEST
This archive was generated by hypermail 2.2.0 : Tue Jan 02 2007 - 14:45:06 MET