From c152c77442de5bc6d34f64bbabf338a726de6a0f Mon Sep 17 00:00:00 2001
From: Maciej Aszyk <maciej@ravendb.net>
Date: Mon, 20 Jan 2025 15:19:29 +0100
Subject: [PATCH] RavenDB-23631 Remove duplicates when performing `AndWith`
 with `TermReader`

---
 src/Corax/Querying/Matches/MultiTermMatch.cs | 10 +++-
 test/SlowTests/Corax/RavenDB_23631.cs        | 57 ++++++++++++++++++++
 2 files changed, 65 insertions(+), 2 deletions(-)
 create mode 100644 test/SlowTests/Corax/RavenDB_23631.cs

diff --git a/src/Corax/Querying/Matches/MultiTermMatch.cs b/src/Corax/Querying/Matches/MultiTermMatch.cs
index a650263fdb7c..67e9aa9c207e 100644
--- a/src/Corax/Querying/Matches/MultiTermMatch.cs
+++ b/src/Corax/Querying/Matches/MultiTermMatch.cs
@@ -367,12 +367,14 @@ private int AndWithFill(Span<long> buffer, int matches)
             using var _ = _context.Allocate(3 * sizeof(long) * buffer.Length, out var bufferHolder);
             var longBuffer = MemoryMarshal.Cast<byte, long>(bufferHolder.ToSpan());
             _termReader.Reset(ref this);
+            
             Span<long> results = longBuffer.Slice(0, buffer.Length);
             Span<long> incomingMatches = longBuffer.Slice(buffer.Length, buffer.Length);
             Span<long> localMatches = longBuffer.Slice(2 * buffer.Length, buffer.Length);
 
             var actualMatches = buffer.Slice(0, matches);
             actualMatches.CopyTo(incomingMatches);
+            
             var currentMatchCount = 0;
             _totalResults = 0;
             
@@ -385,7 +387,11 @@ private int AndWithFill(Span<long> buffer, int matches)
                 fillCounter++;
                 _token.ThrowIfCancellationRequested();
                 _totalResults += read;
-                var common = MergeHelper.And(results, localMatches.Slice(0, read), incomingMatches.Slice(0, matches));
+                var common = MergeHelper.And(
+                    dst: results, 
+                    left: localMatches.Slice(0, read),
+                    right: incomingMatches.Slice(0, matches));
+                
                 results = results.Slice(common);
                 currentMatchCount += common;
             }
@@ -393,7 +399,7 @@ private int AndWithFill(Span<long> buffer, int matches)
             longBuffer.Slice(0, currentMatchCount).CopyTo(buffer);
             
             if (fillCounter > 1)
-                Sort.Run(buffer.Slice(0, currentMatchCount));
+                currentMatchCount = Sorting.SortAndRemoveDuplicates(buffer[..currentMatchCount]);
             
             return currentMatchCount;
         }
diff --git a/test/SlowTests/Corax/RavenDB_23631.cs b/test/SlowTests/Corax/RavenDB_23631.cs
new file mode 100644
index 000000000000..57c25e13bfab
--- /dev/null
+++ b/test/SlowTests/Corax/RavenDB_23631.cs
@@ -0,0 +1,57 @@
+using System;
+using System.Collections.Generic;
+using Corax;
+using Corax.Indexing;
+using Corax.Mappings;
+using Corax.Querying;
+using FastTests.Voron;
+using Sparrow;
+using Tests.Infrastructure;
+using Xunit;
+using Xunit.Abstractions;
+
+namespace SlowTests.Corax;
+
+public class RavenDB_23631(ITestOutputHelper output) : StorageTest(output)
+{
+    [RavenFact(RavenTestCategory.Querying | RavenTestCategory.Corax)]
+    public void MultiTermMatchDoesNotReturnDuplicatesWhenPerformingAndWith()
+    {
+        using var mapping = IndexFieldsMappingBuilder.CreateForWriter(false)
+            .AddBinding(0, "id()")
+            .AddBinding(1, "name")
+            .Build();
+        
+        using (var writer = new IndexWriter(Env, mapping, SupportedFeatures.All))
+        {
+            for (int i = 0; i < 1000; i++)
+            {
+                using (var builder = writer.Index($"id/{i}"))
+                {
+                    builder.Write(0, Encodings.Utf8.GetBytes($"id/{i}"));
+                    builder.IncrementList();
+                    builder.Write(1, Encodings.Utf8.GetBytes("name/0"));
+                    builder.Write(1, Encodings.Utf8.GetBytes("name/1"));
+                    builder.DecrementList();
+                    builder.EndWriting();
+                }
+            }
+            
+            writer.Commit();
+        }
+
+        using (var searcher = new IndexSearcher(Env, mapping))
+        {
+            var @in = searcher.InQuery("id()", ["id/0", "id/10"]);
+            var mtm = searcher.ExistsQuery(mapping.GetByFieldId(1).Metadata);
+
+            var resultMatch = searcher.And(@in, mtm);
+            Span<long> ids = stackalloc long[16];
+            var read = resultMatch.Fill(ids);
+            Assert.Distinct(ids[..read].ToArray());
+            Assert.Equal(2, read);
+            var nothingLeft = resultMatch.Fill(ids) == 0;
+            Assert.True(nothingLeft);
+        }
+    }
+}