[ntuple] Add a join processing tutorial #17137
Conversation
enirolf
requested review from
jblomer,
hahnjo,
pcanal,
silverweed and
vepadulano
|
The failing clang-format is not actually a formatting failure (not sure why the job failed..). I ran clang-format locally before pushing, so that should all be ok 🤷♀️. |
Test Results 18 files 18 suites 4d 2h 42m 38s ⏱️ For more details on these failures, see this check. Results for commit 2d11817. ♻️ This comment has been updated with latest results. |
| /// \macro_code | ||
| /// | ||
| /// \date November 2024 | ||
| /// \author The ROOT Team |
There was a problem hiding this comment.
I would say your name is better suited here
There was a problem hiding this comment.
All other tutorials also have "The ROOT team", in this case I would prefer to keep this consistency.
| auto ntuple = RNTupleWriter::Recreate(std::move(model), ntupleName, ntupleFileName); | ||
|
|
||
| // The main ntuple only contains a subset of the entries present in the auxiliary ntuple. | ||
| for (int i = 0; i < kNEvents; i += gRandom->Integer(5)) { |
There was a problem hiding this comment.
I would say we don't want to have a random number of entries in a tutorial?
| // Fields from the main ntuple are accessed by their original name. | ||
| px = *entry.GetPtr<float>("vpx"); | ||
| // Fields from auxiliary ntuples are accessed by prepending the name of the auxiliary ntuple. | ||
| py = *entry.GetPtr<float>(kAuxNTupleName + ".vpy"); |
There was a problem hiding this comment.
But in this case it could also be accessed without the auxiliary name?
There was a problem hiding this comment.
No; in the current implementation fields in auxiliary fields always have to be prefixed even if they are unique. Apart from the fact that it would be non-trivial to change this I'm also not sure if we would even want that, I think it would only lead to unwanted ambiguities.
There was a problem hiding this comment.
Ok, this is a different behaviour than TTree. Not wrong per se, but we can keep it in mind if discussion arises at some point
A tutorial for join-based processing is currently provided in `ntpl015_processor_join.C`.
enirolf
force-pushed
the
ntuple-join-tutorial
branch
from
31f74e0 to
2d11817
Compare
This PR adds a tutorial for using the
RNTupleProcessorto join two RNTuples. Once RNTupleProcessors are fully composable (see #17132), we might consider merging it withntpl012_processor_chain.C.