Does the order in Filter and Define within rdataframe matter?

ROOT
1

I am applying some filters and defining some variables. My question is if the order matter concerning the result?

For example:
Is rdf.Filter(good_pt.size()>100, “event_cut”).Define(“good_pt”, pt[cuts]) the same as rdf.Define(“good_pt”, pt[cuts]).Filter(good_pt.size()>100, “event_cut”) ??

Please read tips for efficient and successful posting and posting code

ROOT Version: Not Provided
Platform: Not Provided
Compiler: Not Provided

2

Hi @imanol097 ,
as the Filter uses the good_pt column, you have to Define it before. With a recent enough ROOT version you will get an “unknown column” error otherwise.

Cheers,
Enrico

3

And if I do rdf.Define(“good_pt”, pt[cuts]).Filter(good_pt.size()>100, “event_cut”) , then the good_pt variable would have the Filter applied?

4

I am not sure I understand the question. Filter does not filter variables, it filters events. After the Filter call, only events that satisfy good_pt.size() > 100 will be processed.

5

My question is if the “good_pt” would have only events with good_pt.size() > 100

6

good_pt will be calculated for every event, as it’s needed by the Filter. Then the Filter will be evaluated and if good_pt.size() > 100 whatever you put after the Filter will execute, otherwise it will not.

EDIT: more in detail, Defines are evaluated at most once per event, only if something needs that value (in this case, the Filter).

7

What should I do to get good_pt evaluated only in the events with good_pt.size() > 100?

8

I’m probably missing something: leaving aside Define and Filter for a second, how can you evaluate good_pt.size() > 100 without first evaluating good_pt? Seems impossible, independently of RDataFrame.

What am I missing?

9

You are not missing anything. Don’t worry :slight_smile:
Maybe the best option is to define another variable, thanks for your help.

10

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.