[ntuple] optimize bulk read for array of PoD

Author: jblomer

tree/ntuple/inc/ROOT/RField/RFieldSequenceContainer.hxx

@@ -70,6 +70,7 @@ protected:
    std::size_t AppendImpl(const void *from) final;
    void ReadGlobalImpl(ROOT::NTupleSize_t globalIndex, void *to) final;
    void ReadInClusterImpl(RNTupleLocalIndex localIndex, void *to) final;
+   std::size_t ReadBulkImpl(const RBulkSpec &bulkSpec) final;
 
    void ReconcileOnDiskField(const RNTupleDescriptor &desc) final;
 

tree/ntuple/src/RFieldSequenceContainer.cxx

@@ -75,6 +75,16 @@ void ROOT::RArrayField::ReadInClusterImpl(RNTupleLocalIndex localIndex, void *to
    }
 }
 
+std::size_t ROOT::RArrayField::ReadBulkImpl(const RBulkSpec &bulkSpec)
+{
+   if (!fSubfields[0]->IsSimple())
+      return RFieldBase::ReadBulkImpl(bulkSpec);
+
+   GetPrincipalColumnOf(*fSubfields[0])
+      ->ReadV(bulkSpec.fFirstIndex * fArrayLength, bulkSpec.fCount * fArrayLength, bulkSpec.fValues);
+   return RBulkSpec::kAllSet;
+}
+
 void ROOT::RArrayField::ReconcileOnDiskField(const RNTupleDescriptor &desc)
 {
    static const std::vector<std::string> prefixes = {"std::array<"};

tree/ntuple/test/ntuple_bulk.cxx

@@ -208,6 +208,55 @@ TEST(RNTupleBulk, RVec)
    }
 }
 
+TEST(RNTupleBulk, Array)
+{
+   FileRaii fileGuard("test_ntuple_bulk_array.root");
+   {
+      auto model = RNTupleModel::Create();
+      auto fldArrI = model->MakeField<std::array<int, 2>>("aint");
+      auto fld2DArrI = model->MakeField<std::array<std::array<int, 2>, 2>>("2daint");
+      auto writer = RNTupleWriter::Recreate(std::move(model), "ntpl", fileGuard.GetPath());
+      for (int i = 0; i < 3; ++i) {
+         fldArrI->at(0) = 2 * i;
+         fldArrI->at(1) = 2 * i + 1;
+
+         fld2DArrI->at(0).at(0) = 4 * i;
+         fld2DArrI->at(0).at(1) = 4 * i + 1;
+         fld2DArrI->at(1).at(0) = 4 * i + 2;
+         fld2DArrI->at(1).at(1) = 4 * i + 3;
+
+         writer->Fill();
+      }
+   }
+
+   auto reader = RNTupleReader::Open("ntpl", fileGuard.GetPath());
+   const auto &model = reader->GetModel();
+
+   RFieldBase::RBulkValues bulkI = model.CreateBulk("aint");
+   RFieldBase::RBulkValues bulk2DI = model.CreateBulk("2daint");
+
+   auto mask = std::make_unique<bool[]>(3);
+   mask[0] = true;
+   mask[1] = false; // the std::array<simple type, ...> field optimization should ignore the mask
+   mask[2] = true;
+
+   auto iArr = static_cast<std::array<int, 2> *>(bulkI.ReadBulk(RNTupleLocalIndex(0, 0), mask.get(), 3));
+   auto i2DArr =
+      static_cast<std::array<std::array<int, 2>, 2> *>(bulk2DI.ReadBulk(RNTupleLocalIndex(0, 0), mask.get(), 3));
+
+   for (int i = 0; i < 3; ++i) {
+      EXPECT_EQ(2 * i, iArr[i][0]);
+      EXPECT_EQ(2 * i + 1, iArr[i][1]);
+
+      if (mask[i]) {
+         EXPECT_EQ(4 * i, i2DArr[i][0][0]);
+         EXPECT_EQ(4 * i + 1, i2DArr[i][0][1]);
+         EXPECT_EQ(4 * i + 2, i2DArr[i][1][0]);
+         EXPECT_EQ(4 * i + 3, i2DArr[i][1][1]);
+      }
+   }
+}
+
 TEST(RNTupleBulk, Adopted)
 {
    FileRaii fileGuard("test_ntuple_bulk_adopted.root");