[ntuple] Add RPageSource::OpenWithDifferentAnchor

tree/ntuple/inc/ROOT/RMiniFile.hxx

@@ -32,13 +32,12 @@ class TVirtualStreamerInfo;
 
 namespace ROOT {
 
-namespace Internal {
-class RRawFile;
-}
-
 class RNTupleWriteOptions;
 
 namespace Internal {
+
+class RRawFile;
+
 /// Holds status information of an open ROOT file during writing
 struct RTFileControlBlock;
 
@@ -68,9 +67,6 @@ private:
    /// Used when the file turns out to be a TFile container. The ntuplePath variable is either the ntuple name
    /// or an ntuple name preceded by a directory (`myNtuple` or `foo/bar/myNtuple` or `/foo/bar/myNtuple`)
    RResult<RNTuple> GetNTupleProper(std::string_view ntuplePath);
-   /// Loads an RNTuple anchor from a TFile at the given file offset (unzipping it if necessary).
-   RResult<RNTuple>
-   GetNTupleProperAtOffset(std::uint64_t payloadOffset, std::uint64_t compSize, std::uint64_t uncompLen);
 
    /// Searches for a key with the given name and type in the key index of the directory starting at offsetDir.
    /// The offset points to the start of the TDirectory DATA section, without the key and without the name and title
@@ -84,6 +80,9 @@ public:
    explicit RMiniFileReader(ROOT::Internal::RRawFile *rawFile);
    /// Extracts header and footer location for the RNTuple identified by ntupleName
    RResult<RNTuple> GetNTuple(std::string_view ntupleName);
+   /// Loads an RNTuple anchor from a TFile at the given file offset (unzipping it if necessary).
+   RResult<RNTuple>
+   GetNTupleProperAtOffset(std::uint64_t payloadOffset, std::uint64_t compSize, std::uint64_t uncompLen);
    /// Reads a given byte range from the file into the provided memory buffer.
    /// If `nbytes > fMaxKeySize` it will perform chunked read from multiple blobs,
    /// whose addresses are listed at the end of the first chunk.

tree/ntuple/inc/ROOT/RPageStorage.hxx

@@ -757,7 +757,7 @@ public:
    /// The underlying `std::shared_mutex`, however, is neither read nor write recursive:
    /// within one thread, only one lock (shared or exclusive) must be acquired at the same time. This requires special
    /// care in sections protected by `GetSharedDescriptorGuard()` and `GetExclDescriptorGuard()` especially to avoid
-   /// that the locks are acquired indirectly (e.g. by a call to `GetNEntries()`). As a general guideline, no other
+   /// that the locks are acquired indirectly. As a general guideline, no other
    /// method of the page source should be called (directly or indirectly) in a guarded section.
    const RSharedDescriptorGuard GetSharedDescriptorGuard() const
    {

tree/ntuple/inc/ROOT/RPageStorageDaos.hxx

@@ -77,7 +77,8 @@ struct RDaosNTupleAnchor {
    /// The object class for user data OIDs, e.g. `SX`
    std::string fObjClass{};
 
-   bool operator ==(const RDaosNTupleAnchor &other) const {
+   bool operator==(const RDaosNTupleAnchor &other) const
+   {
       return fVersionAnchor == other.fVersionAnchor && fVersionEpoch == other.fVersionEpoch &&
              fVersionMajor == other.fVersionMajor && fVersionMinor == other.fVersionMinor &&
              fVersionPatch == other.fVersionPatch && fNBytesHeader == other.fNBytesHeader &&

tree/ntuple/inc/ROOT/RPageStorageFile.hxx

@@ -179,6 +179,11 @@ public:
    RPageSourceFile &operator=(RPageSourceFile &&) = delete;
    ~RPageSourceFile() override;
 
+   /// Creates a new PageSourceFile using the same underlying file as this but referring to a different RNTuple,
+   /// represented by `anchor`.
+   std::unique_ptr<RPageSourceFile>
+   OpenWithDifferentAnchor(const RNTuple &anchor, const ROOT::RNTupleReadOptions &options = ROOT::RNTupleReadOptions());
+
    void
    LoadSealedPage(ROOT::DescriptorId_t physicalColumnId, RNTupleLocalIndex localIndex, RSealedPage &sealedPage) final;
 

tree/ntuple/src/RPageStorageFile.cxx

@@ -350,6 +350,15 @@ ROOT::Internal::RPageSourceFile::CreateFromAnchor(const RNTuple &anchor, const R
 
 ROOT::Internal::RPageSourceFile::~RPageSourceFile() = default;
 
+std::unique_ptr<ROOT::Internal::RPageSourceFile>
+ROOT::Internal::RPageSourceFile::OpenWithDifferentAnchor(const RNTuple &anchor, const ROOT::RNTupleReadOptions &options)
+{
+   auto pageSource = std::make_unique<RPageSourceFile>("", fFile->Clone(), options);
+   pageSource->fAnchor = anchor;
+   // NOTE: fNTupleName gets set only upon Attach().
+   return pageSource;
+}
+
 void ROOT::Internal::RPageSourceFile::LoadStructureImpl()
 {
    // If we constructed the page source with (ntuple name, path), we need to find the anchor first.
@@ -552,18 +561,17 @@ ROOT::Internal::RPageSourceFile::PrepareSingleCluster(const RCluster::RKey &clus
    std::vector<ROnDiskPageLocator> onDiskPages;
    auto activeSize = 0;
    auto pageZeroMap = std::make_unique<ROnDiskPageMap>();
-   PrepareLoadCluster(clusterKey, *pageZeroMap,
-                      [&](ROOT::DescriptorId_t physicalColumnId, ROOT::NTupleSize_t pageNo,
-                          const ROOT::RClusterDescriptor::RPageInfo &pageInfo) {
-                         const auto &pageLocator = pageInfo.GetLocator();
-                         if (pageLocator.GetType() == RNTupleLocator::kTypeUnknown)
-                            throw RException(R__FAIL("tried to read a page with an unknown locator"));
-                         const auto nBytes =
-                            pageLocator.GetNBytesOnStorage() + pageInfo.HasChecksum() * kNBytesPageChecksum;
-                         activeSize += nBytes;
-                         onDiskPages.push_back(
-                            {physicalColumnId, pageNo, pageLocator.GetPosition<std::uint64_t>(), nBytes, 0});
-                      });
+   PrepareLoadCluster(
+      clusterKey, *pageZeroMap,
+      [&](ROOT::DescriptorId_t physicalColumnId, ROOT::NTupleSize_t pageNo,
+          const ROOT::RClusterDescriptor::RPageInfo &pageInfo) {
+         const auto &pageLocator = pageInfo.GetLocator();
+         if (pageLocator.GetType() == RNTupleLocator::kTypeUnknown)
+            throw RException(R__FAIL("tried to read a page with an unknown locator"));
+         const auto nBytes = pageLocator.GetNBytesOnStorage() + pageInfo.HasChecksum() * kNBytesPageChecksum;
+         activeSize += nBytes;
+         onDiskPages.push_back({physicalColumnId, pageNo, pageLocator.GetPosition<std::uint64_t>(), nBytes, 0});
+      });
 
    // Linearize the page requests by file offset
    std::sort(onDiskPages.begin(), onDiskPages.end(),

tree/ntuple/test/ntuple_storage.cxx

@@ -1145,3 +1145,62 @@ TEST(RPageSourceFile, NameFromAnchor)
    source->Attach();
    EXPECT_EQ(source->GetNTupleName(), "ntpl");
 }
+
+TEST(RPageSourceFile, OpenDifferentAnchor)
+{
+   FileRaii fileGuard("test_ntuple_open_diff_anchor.root");
+
+   auto model = RNTupleModel::Create();
+   auto pF = model->MakeField<float>("f");
+   auto file = std::unique_ptr<TFile>(TFile::Open(fileGuard.GetPath().c_str(), "RECREATE"));
+   {
+      auto writer = RNTupleWriter::Append(std::move(model), "ntpl1", *file);
+      for (auto i = 0; i < 100; ++i) {
+         *pF = i;
+         writer->Fill();
+      }
+   }
+   {
+      model = RNTupleModel::Create();
+      auto pI = model->MakeField<int>("i");
+      auto pC = model->MakeField<char>("c");
+
+      auto writer = RNTupleWriter::Append(std::move(model), "ntpl2", *file);
+      for (auto i = 0; i < 20; ++i) {
+         *pI = i;
+         *pC = i;
+         writer->Fill();
+      }
+   }
+
+   auto source = std::make_unique<RPageSourceFile>("ntpl1", fileGuard.GetPath(), RNTupleReadOptions());
+   source->Attach();
+   EXPECT_EQ(source->GetNEntries(), 100);
+   {
+      auto desc = source->GetSharedDescriptorGuard();
+      EXPECT_NE(desc->FindFieldId("f"), ROOT::kInvalidDescriptorId);
+   }
+
+   auto anchor2 = file->Get<ROOT::RNTuple>("ntpl2");
+   ASSERT_NE(anchor2, nullptr);
+   auto source2 = source->OpenWithDifferentAnchor(*anchor2);
+   source2->Attach();
+   EXPECT_EQ(source2->GetNTupleName(), "ntpl2");
+   EXPECT_EQ(source2->GetNEntries(), 20);
+   {
+      auto desc2 = source2->GetSharedDescriptorGuard();
+      EXPECT_EQ(desc2->FindFieldId("f"), ROOT::kInvalidDescriptorId);
+      EXPECT_NE(desc2->FindFieldId("i"), ROOT::kInvalidDescriptorId);
+      EXPECT_NE(desc2->FindFieldId("c"), ROOT::kInvalidDescriptorId);
+   }
+
+   source.reset();
+   // source2 should still be valid after dropping the first source.
+   EXPECT_EQ(source2->GetNEntries(), 20);
+   {
+      auto desc2 = source2->GetSharedDescriptorGuard();
+      EXPECT_EQ(desc2->FindFieldId("f"), ROOT::kInvalidDescriptorId);
+      EXPECT_NE(desc2->FindFieldId("i"), ROOT::kInvalidDescriptorId);
+      EXPECT_NE(desc2->FindFieldId("c"), ROOT::kInvalidDescriptorId);
+   }
+}