Removing Duplicate Entries in a ROOT TTree

Question

I have three .root files that I need to merge together. Normally I would use hadd to merge the files, but the files contain duplicate entries which I need to remove. I can't just delete the duplicated entries because TTrees are read-only. Is there a simple way to merge the files while ensuring that only unique entries are saved?

Answer 1

I did manage to find a way to produce histograms that contain only unique entries using TEntryList. This allows you to specify which tree entries you wish to use. In my case, each entry has an event number which identifies it. So, I generated an entry list with the entry numbers corresponding to only unique event numbers.

set<int> eventIds; // keep track of already seen event numbers
int EVENT;
int nEntries = tree->GetEntries();

tree->SetBranchAddress("EVENT",&EVENT); // grab the event number from the tree

TEntryList *tlist = new TEntryList(tree); // initialize entry list for 'TTree* tree'

// loop over the entries in 'tree'
for (int j = 0; j < nEntries; ++j)
{
    tree->GetEvent(j);

    // if we have not seen this event yet, add it to the set
    // and to the entry list
    if (eventIds.count(EVENT) == 0)
    {
        eventIds.insert(EVENT);
        tlist->Enter(j,tree);
    }
}

// apply the entry list to the tree
tree->SetEntryList(tlist);

// histogram of the variable 'var' will be drawn only with the
// entries specified by the entry list.
tree->Draw("var");