From 2d3891628f707d8f2d4a32cd77b2749746705bb9 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 88bf518fd6130..96c67dc29055f 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 c8ff26b3bcd672b758c6844392958f766e7fd94c 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      | 48 +++++++++++++++++++
 2 files changed, 52 insertions(+)

diff --git a/tree/ntupleutil/inc/ROOT/RNTupleInspector.hxx b/tree/ntupleutil/inc/ROOT/RNTupleInspector.hxx
index 96c67dc29055f..384a3dc9bab50 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 histName = "", std::string histTitle = "",
+                                                           size_t nBins = 64);
+
    /////////////////////////////////////////////////////////////////////////////
    /// \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 f906bbf523a6d..4153ce40e526f 100644
--- a/tree/ntupleutil/src/RNTupleInspector.cxx
+++ b/tree/ntupleutil/src/RNTupleInspector.cxx
@@ -455,6 +455,54 @@ 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 histName, std::string histTitle, size_t nBins)
+{
+   if (histName.empty())
+      histName = "pagesPerClusterHist";
+   if (histTitle.empty())
+      histTitle = "Per-column type pages per cluster distribution";
+
+   auto stackedHist = std::make_unique<THStack>(histName.c_str(), histTitle.c_str());
+
+   double histMin = std::numeric_limits<double>::max();
+   double histMax = 0;
+   std::vector<std::unordered_map<ENTupleColumnType, std::uint64_t>> pagesPerCluster;
+
+   std::vector<ROOT::ENTupleColumnType> colTypeVec = colTypes;
+   if (std::empty(colTypes)) {
+      colTypeVec = GetColumnTypes();
+   }
+
+   for (const auto &clDesc : fDescriptor.GetClusterIterable()) {
+      auto &pagesInThisCluster = pagesPerCluster.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 &pagesPerType = pagesInThisCluster[colDesc.GetType()];
+         pagesPerType += nPages;
+         histMin = std::min(histMin, static_cast<double>(pagesPerType));
+         histMax = std::max(histMax, static_cast<double>(pagesPerType));
+      }
+   }
+
+   for (const auto colType : colTypeVec) {
+      auto hist = std::make_unique<TH1D>(
+         TString::Format("%s%s", histName.c_str(), RColumnElementBase::GetColumnTypeName(colType)),
+         RColumnElementBase::GetColumnTypeName(colType), nBins, histMin,
+         histMax + ((histMax - histMin) / static_cast<double>(nBins)));
+      for (const auto &pagesInCluster : pagesPerCluster) {
+         auto it = pagesInCluster.find(colType);
+         hist->Fill(it == pagesInCluster.end() ? 0u : it->second);
+      }
+
+      stackedHist->Add(hist.release());
+   }
+
+   return stackedHist;
+}
+
 //------------------------------------------------------------------------------
 
 const ROOT::Experimental::RNTupleInspector::RFieldTreeInspector &