Segfault when defining new column with gInterpreter

[jcob]

Hi,

I’m working in a python environment but I’m declaring some C++ functions with gInterpreter to create some new columns/branches in an RDataFrame. I do this with DisableImplicitMT().

This function is as follows and defined using ROOT.gInterpreter.Declare:

std::tuple<ROOT::RDF::RNode,ROOT::RDF::RNode> ws_rs_contamination(ROOT::RDataFrame ws_df, ROOT::RDataFrame rs_df){
    ROOT::RDF::RNode ws_node = ws_df;
    ROOT::RDF::RNode rs_node = rs_df;
    ws_node = ws_node.Define("unique_event_number",[](const unsigned int& rn, const unsigned long long& en) -> unsigned long long {return 10e+9*rn+en;},{"runNumber","eventNumber"});
    rs_node = rs_node.Define("unique_event_number",[](const unsigned int& rn, const unsigned long long& en) -> unsigned long long {return 10e+9*rn+en;},{"runNumber","eventNumber"});
    std::vector<unsigned long long> uevs_ws;
    std::vector<unsigned long long> uevs_rs;
    
    ws_node.Foreach([&uevs_ws](const unsigned long long& uev){
    uevs_ws.push_back(uev);
    }, {"unique_event_number"});
    std::cout << "Length of WS uev vector: " << uevs_ws.size() << std::endl;
    
    rs_node.Foreach([&uevs_rs](const unsigned long long& uev){
    uevs_rs.push_back(uev);
    }, {"unique_event_number"});
    std::cout << "Length of RS uev vector: " << uevs_rs.size() << std::endl;
    
    //sort the uev vectors
    std::sort(uevs_ws.begin(), uevs_ws.end());
    std::sort(uevs_rs.begin(), uevs_rs.end());
    
    //find unique event numbers that are in both the RS and WS.
    std::vector<unsigned long long> common_uevs;
    std::set_intersection(uevs_ws.begin(), uevs_ws.end(), uevs_rs.begin(), uevs_rs.end(), std::back_inserter(common_uevs));
    ws_node = ws_node.Define("in_both",[&common_uevs] (unsigned long long uev) {return (bool)std::binary_search(common_uevs.begin(),common_uevs.end(),uev);},{"unique_event_number"});
    rs_node = rs_node.Define("in_both",[&common_uevs] (unsigned long long uev) {return (bool)std::binary_search(common_uevs.begin(),common_uevs.end(),uev);},{"unique_event_number"});
    return {ws_node,rs_node};

When I return the dataframes using this function and try the following:

ws_test,rs_test = ROOT.ws_rs_contamination(ws_frame_no_clones,rs_frame_no_clones)
arr = ws_test.AsNumpy(["in_both"])

I get a segfault with the error pointing to this

     69     result_ptrs = {}
     70     for column in columns:
---> 71         column_type = df.GetColumnType(column)
     72         result_ptrs[column] = df.Take[column_type](column)

Is this an issue with gInterpreter and python or am I doing something wrong in my code?

Thanks

---

Please read tips for efficient and successful posting and posting code

_ROOT Version: v6.22

---

[eguiraud]

Hi @jcob ,
can you please post the full error output? Also can you check whether this still fails with a nightly build of ROOT?

Cheers,
Enrico

[jcob]

Hi @eguiraud,

The full error output is here:

     69     result_ptrs = {}
     70     for column in columns:
---> 71         column_type = df.GetColumnType(column)
     72         result_ptrs[column] = df.Take[column_type](column)
     73 

SegmentationViolation: string ROOT::RDF::RInterface<ROOT::Detail::RDF::RNodeBase,void>::GetColumnType(basic_string_view<char,char_traits<char> > column) =>
    SegmentationViolation: segfault in C++; program state was reset

I’ll try with a nightly build.

Thanks

[eguiraud]

Wow that’s a very short stacktrace, quite surprising – a lot of context is missing.
If the nightly builds also behave like that, I think we’ll need a self-contained reproducer that we can run on our side, together with sample data (just a few events should be enough). I don’t know how else to debug this (at first glance there is nothing wrong in your code).

Cheers,
Enrico

[jcob]

Hi @eguiraud,

I tried re-running this with the root nightlies (via conda) and it throws the same error. I’ll try and produce a reproducer for this.

Thanks