From 9412b5debd61f7854068fb715f716449508ded30 Mon Sep 17 00:00:00 2001
From: Igor Markelov <ismarkelov@edu.hse.ru>
Date: Sun, 26 May 2024 19:23:17 +0000
Subject: [PATCH] Add estimateCardinalityInPermutedRange for ColumnVector

---
 src/Columns/ColumnVector.cpp | 20 ++++++++++++++++++++
 src/Columns/ColumnVector.h   |  2 ++
 2 files changed, 22 insertions(+)

diff --git a/src/Columns/ColumnVector.cpp b/src/Columns/ColumnVector.cpp
index 4e3b9963107..498b9cb7c32 100644
--- a/src/Columns/ColumnVector.cpp
+++ b/src/Columns/ColumnVector.cpp
@@ -14,6 +14,7 @@
 #include <Common/Arena.h>
 #include <Common/Exception.h>
 #include <Common/HashTable/Hash.h>
+#include <Common/HashTable/StringHashSet.h>
 #include <Common/NaNUtils.h>
 #include <Common/RadixSort.h>
 #include <Common/SipHash.h>
@@ -413,6 +414,25 @@ void ColumnVector<T>::updatePermutation(IColumn::PermutationSortDirection direct
     }
 }
 
+template<typename T>
+size_t ColumnVector<T>::estimateCardinalityInPermutedRange(const IColumn::Permutation & permutation, const EqualRange & equal_range) const
+{
+    const size_t range_size = equal_range.size();
+    if (range_size <= 1)
+        return range_size;
+
+    /// TODO use sampling if the range is too large (e.g. 16k elements, but configurable)
+    StringHashSet elements;
+    bool inserted = false;
+    for (size_t i = equal_range.from; i < equal_range.to; ++i)
+    {
+        size_t id = permutation[i];
+        StringRef ref = getDataAt(id);
+        elements.emplace(ref, inserted);
+    }
+    return elements.size();
+}
+
 template <typename T>
 MutableColumnPtr ColumnVector<T>::cloneResized(size_t size) const
 {
diff --git a/src/Columns/ColumnVector.h b/src/Columns/ColumnVector.h
index 91bceaa4534..bbd27c91a70 100644
--- a/src/Columns/ColumnVector.h
+++ b/src/Columns/ColumnVector.h
@@ -161,6 +161,8 @@ public:
     void updatePermutation(IColumn::PermutationSortDirection direction, IColumn::PermutationSortStability stability,
                     size_t limit, int nan_direction_hint, IColumn::Permutation & res, EqualRanges& equal_ranges) const override;
 
+    size_t estimateCardinalityInPermutedRange(const IColumn::Permutation & permutation, const EqualRange & equal_range) const override;
+
     void reserve(size_t n) override
     {
         data.reserve_exact(n);