From 6470738260c9f55ca7fe5ad5e74c1a1b9ad79ccc Mon Sep 17 00:00:00 2001
From: silverweed <giacomo.parolini@cern.ch>
Date: Thu, 3 Sep 2026 14:45:09 +0200
Subject: [PATCH 1/2] [ntupleutil] clang-format RNTupleInspector.hxx

---
 tree/ntupleutil/inc/ROOT/RNTupleInspector.hxx | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/tree/ntupleutil/inc/ROOT/RNTupleInspector.hxx b/tree/ntupleutil/inc/ROOT/RNTupleInspector.hxx
index 957b1550b9c88..30bc3ce254f96 100644
--- a/tree/ntupleutil/inc/ROOT/RNTupleInspector.hxx
+++ b/tree/ntupleutil/inc/ROOT/RNTupleInspector.hxx
@@ -105,7 +105,9 @@ public:
          : fColumnDescriptor(colDesc),
            fCompressedPageSizes(compressedPageSizes),
            fElementSize(elemSize),
-           fNElements(nElems){};
+           fNElements(nElems)
+      {
+      }
       ~RColumnInspector() = default;
 
       const ROOT::RColumnDescriptor &GetDescriptor() const { return fColumnDescriptor; }
@@ -136,7 +138,9 @@ public:
 
    public:
       RFieldTreeInspector(const ROOT::RFieldDescriptor &fieldDesc, std::uint64_t onDiskSize, std::uint64_t inMemSize)
-         : fRootFieldDescriptor(fieldDesc), fCompressedSize(onDiskSize), fUncompressedSize(inMemSize){};
+         : fRootFieldDescriptor(fieldDesc), fCompressedSize(onDiskSize), fUncompressedSize(inMemSize)
+      {
+      }
       ~RFieldTreeInspector() = default;
 
       const ROOT::RFieldDescriptor &GetDescriptor() const { return fRootFieldDescriptor; }

From 363ba670134159a78e7eebbe8d719c04f33baead Mon Sep 17 00:00:00 2001
From: silverweed <giacomo.parolini@cern.ch>
Date: Thu, 3 Sep 2026 14:39:21 +0200
Subject: [PATCH 2/2] [ntupleutil] add
 RNTupleInspector::GetPagesPerClusterDistribution

---
 tree/ntupleutil/inc/ROOT/RNTupleInspector.hxx |  4 ++
 tree/ntupleutil/src/RNTupleInspector.cxx      | 64 +++++++++++++++++++
 2 files changed, 68 insertions(+)

diff --git a/tree/ntupleutil/inc/ROOT/RNTupleInspector.hxx b/tree/ntupleutil/inc/ROOT/RNTupleInspector.hxx
index 30bc3ce254f96..407c58c408bfa 100644
--- a/tree/ntupleutil/inc/ROOT/RNTupleInspector.hxx
+++ b/tree/ntupleutil/inc/ROOT/RNTupleInspector.hxx
@@ -430,6 +430,10 @@ public:
                                                     std::string histName = "", std::string histTitle = "",
                                                     size_t nBins = 64);
 
+   std::unique_ptr<THStack> GetPagesPerClusterDistribution(std::initializer_list<ROOT::ENTupleColumnType> colTypes = {},
+                                                           std::string_view histName = "",
+                                                           std::string_view histTitle = "", size_t nBins = 32);
+
    /////////////////////////////////////////////////////////////////////////////
    /// \brief Get storage information for a given (sub)field by ID.
    ///
diff --git a/tree/ntupleutil/src/RNTupleInspector.cxx b/tree/ntupleutil/src/RNTupleInspector.cxx
index a3287a1d4d2ff..5b0bf5154e0d4 100644
--- a/tree/ntupleutil/src/RNTupleInspector.cxx
+++ b/tree/ntupleutil/src/RNTupleInspector.cxx
@@ -455,6 +455,70 @@ ROOT::Experimental::RNTupleInspector::GetPageSizeDistribution(std::initializer_l
    return stackedHist;
 }
 
+std::unique_ptr<THStack> ROOT::Experimental::RNTupleInspector::GetPagesPerClusterDistribution(
+   std::initializer_list<ROOT::ENTupleColumnType> colTypes, std::string_view histName, std::string_view histTitle,
+   size_t nBins)
+{
+   if (histName.empty())
+      histName = "pagesPerClusterHist";
+   if (histTitle.empty())
+      histTitle = "#pages per cluster";
+
+   auto stackedHist = std::make_unique<THStack>(std::string(histName).c_str(), std::string(histTitle).c_str());
+   // Each element of the vector contains the sum of all pages of all columns of each type.
+   std::vector<std::unordered_map<ENTupleColumnType, std::uint64_t>> pagesPerClusterPerColumnType;
+
+   std::vector<ROOT::ENTupleColumnType> colTypeVec = colTypes;
+   if (std::empty(colTypes)) {
+      colTypeVec = GetColumnTypes();
+   }
+
+   for (const auto &clDesc : fDescriptor.GetClusterIterable()) {
+      auto &pagesPerColumnTypeInThisCluster = pagesPerClusterPerColumnType.emplace_back();
+      for (const auto &colRange : clDesc.GetColumnRangeIterable()) {
+         const auto colId = colRange.GetPhysicalColumnId();
+         const auto &colDesc = fDescriptor.GetColumnDescriptor(colId);
+         const auto nPages = clDesc.GetPageRange(colId).GetPageInfos().size();
+         auto &nPagesForThisType = pagesPerColumnTypeInThisCluster[colDesc.GetType()];
+         nPagesForThisType += nPages;
+      }
+   }
+
+   double histMin = std::numeric_limits<double>::max();
+   double histMax = 0;
+   for (const auto &pagesPerType : pagesPerClusterPerColumnType) {
+      auto [min, max] = std::minmax_element(pagesPerType.begin(), pagesPerType.end(),
+                                            [](const auto &a, const auto &b) { return a.second < b.second; });
+      histMin = std::min<double>(histMin, min->second);
+      histMax = std::max<double>(histMax, max->second);
+   }
+
+   std::cout << "histMin = " << histMin << ", histMax = " << histMax << "\n";
+
+   std::array<std::unique_ptr<TH1L>, static_cast<std::size_t>(ENTupleColumnType::kMax)> histsPerType;
+   for (const auto &pagesPerType : pagesPerClusterPerColumnType) {
+      for (const auto &[colType, nPages] : pagesPerType) {
+         auto &hist = histsPerType[static_cast<std::size_t>(colType)];
+         if (!hist) {
+            hist =
+               std::make_unique<TH1L>(ROOT::Internal::RColumnElementBase::GetColumnTypeName(colType),
+                                      TString::Format("%s_%s", std::string(histTitle).c_str(),
+                                                      ROOT::Internal::RColumnElementBase::GetColumnTypeName(colType)),
+                                      nBins, histMin, histMax + ((histMax - histMin) / static_cast<double>(nBins)));
+         }
+         hist->Fill(nPages);
+      }
+   }
+
+   for (auto &hist : histsPerType) {
+      if (hist) {
+         stackedHist->Add(hist.release());
+      }
+   }
+
+   return stackedHist;
+}
+
 //------------------------------------------------------------------------------
 
 const ROOT::Experimental::RNTupleInspector::RFieldTreeInspector &