How to apply two cut with in .arrays?

[ZhenxuanZhang-Jensen]

Hello everyone:
I am trying to do some cuts in the .arrays, but no any description can do that with more than two cut apply.
for example:
events_250 = uproot.open("M250.root:Events")
and then apply two cuts on it:
events_250.arrays(cut = ("GEN_deltaR_MAX4Q>1.2" and "GEN_deltaR_MIN4Q<0.8"),library ="np")["GEN_deltaR_MIN4Q"]

[jpivarski]

You can write cut = "(GEN_deltaR_MAX4Q>1.2) & (GEN_deltaR_MIN4Q<0.8)" (with parentheses!), but also, there's no strong reason to put the cut in this string. You can cut the arrays after the .arrays call just as easily:

ttree = uproot.open("M250.root:Events")
one_branch = ttree["GEN_deltaR_MIN4Q"]
GEN_deltaR_MAX4Q = one_branch.array(library ="np")
GEN_deltaR_MAX4Q_cut = GEN_deltaR_MAX4Q[(GEN_deltaR_MAX4Q>1.2) & (GEN_deltaR_MIN4Q<0.8)]

The cut argument does nothing more than put the string you give it into square brackets, which is how you cut in NumPy.

Also, calling .arrays without specifying any expressions or filter_branches reads all branches from disk. Putting ["GEN_deltaR_MIN4Q"] after it picks out one of the already-read branches and throws away the rest. .array (above) is the easiest way to read only one branch, but you can also use .arrays with expressions or filter_branches to read a selection of branches.

It might not be relevant on small files, but on big files, reading unnecessary branches can be slow or impossible of you don't have enough memory.

[ZhenxuanZhang-Jensen]

Thank you for your answer. I still have confusion about how you apply the cut to all_events. it seems all_events is not even a variable or anything. The other thing is about the .array , you use one_branch.array only read one branch without any expressions?

[jpivarski]

Sorry, I fixed it above. I was trying to write it on my phone this morning, and I changed the code a few times for clarity.

In the above version, one_branch is a branch object and GEN_deltaR_MAX4Q is a NumPy array. If you pull a TBranch out of a TTree (one_branch), you can call TBranch.array on it to get one array. If you don't, you can call TTree.arrays on the TTree to get multiple arrays. With library="np", it's a Python dict of NumPy arrays; with library="ak" (default), it's a single Awkward Array (with record structure instead of a dict, which is a bit more flexible); with library="pd", it's a single Pandas DataFrame. See the links above for details.

But as for your original question, to get the "logical and" of multiple cuts, use the & operator with parentheses. (For the importance of parentheses, see #508.) It does not matter whether you do that inside the cut string or in Python after the array has been loaded. (A lot of people have been thinking that there's something special about putting the cut in the cut string and have been going to great lengths to express difficult things in it, when they could more easily be done in multiple lines of Python code. I want everyone to know that the only advantage cut provides is the convenience of putting the code inside square brackets for you!)