Support prefix/suffix early exit

Author: Genbox

Src/FastData.Benchmarks/Benchmarks/EarlyExitVsLookup.cs

@@ -0,0 +1,48 @@
+namespace Genbox.FastData.Benchmarks.Benchmarks;
+
+/// <summary>Benchmarks where the threshold usually is between a full hash/lookup in a HashSet vs. how long a prefix/suffix early exit can be before it no longer makes sense</summary>
+public class EarlyExitVsLookup
+{
+    private static string _myStr = "hello world";
+
+    private static readonly HashSet<string> _hashSet = new HashSet<string>(StringComparer.Ordinal)
+    {
+        _myStr
+    };
+
+    [Benchmark(Baseline = true)]
+    public bool Lookup() => _hashSet.Contains(_myStr);
+
+    [Benchmark]
+    public bool Size1() => _myStr.StartsWith('h');
+
+    [Benchmark]
+    public bool Size2() => _myStr.StartsWith("he", StringComparison.Ordinal);
+
+    [Benchmark]
+    public bool Size3() => _myStr.StartsWith("hel", StringComparison.Ordinal);
+
+    [Benchmark]
+    public bool Size4() => _myStr.StartsWith("hell", StringComparison.Ordinal);
+
+    [Benchmark]
+    public bool Size5() => _myStr.StartsWith("hello", StringComparison.Ordinal);
+
+    [Benchmark]
+    public bool Size6() => _myStr.StartsWith("hello ", StringComparison.Ordinal);
+
+    [Benchmark]
+    public bool Size7() => _myStr.StartsWith("hello w", StringComparison.Ordinal);
+
+    [Benchmark]
+    public bool Size8() => _myStr.StartsWith("hello wo", StringComparison.Ordinal);
+
+    [Benchmark]
+    public bool Size9() => _myStr.StartsWith("hello wor", StringComparison.Ordinal);
+
+    [Benchmark]
+    public bool Size10() => _myStr.StartsWith("hello worl", StringComparison.Ordinal);
+
+    [Benchmark]
+    public bool Size11() => _myStr.StartsWith("hello world", StringComparison.Ordinal);
+}

Src/FastData.Cli.Tests/CommandOutputs/cpp_-k UInt8_Files_Integers.input.verified.txt

@@ -14,6 +14,7 @@ public:
     static constexpr bool contains(const uint8_t key) noexcept {
         if ((static_cast<uint8_t>(key) & 129) != 0)
             return false;
+
         if (key == 42 || key == 62 || key == 80 || key == 100)
             return true;
 

Src/FastData.Cli.Tests/CommandOutputs/cpp_-s HashTable_Files_Strings.input.verified.txt

@@ -39,8 +39,9 @@ class MyData final {
 public:
     [[nodiscard]]
     static constexpr bool contains(const std::string_view key) noexcept {
-        if (key.length() != 5u)
-            return false;
+         if (key.length() != 5u)
+             return false;
+
 
         const uint64_t hash = get_hash(key);
         const size_t index = hash % 2;

Src/FastData.Cli.Tests/CommandOutputs/cpp_Files_Strings.input.verified.txt

@@ -12,8 +12,9 @@ class MyData final {
 public:
     [[nodiscard]]
     static constexpr bool contains(const std::string_view key) noexcept {
-        if (key.length() != 5u)
-            return false;
+         if (key.length() != 5u)
+             return false;
+
         if (key == "test1" || key == "test2")
             return true;
 

Src/FastData.Cli.Tests/CommandOutputs/csharp_-k UInt8_Files_Integers.input.verified.txt

@@ -16,6 +16,7 @@ internal static class MyData
         if (((byte)key & 129) != 0)
             return false;
 
+
         switch (key)
         {
             case 42:

Src/FastData.Cli.Tests/CommandOutputs/csharp_-s HashTable_Files_Strings.input.verified.txt

@@ -61,6 +61,7 @@ internal static class MyData
         if (key.Length != 5u)
             return false;
 
+
         ulong hash = Hash(key);
         uint index = (uint)(hash & 1);
         sbyte i = (sbyte)(_buckets[index] - 1);

Src/FastData.Cli.Tests/CommandOutputs/csharp_Files_Strings.input.verified.txt

@@ -16,6 +16,7 @@ internal static class MyData
         if (key.Length != 5u)
             return false;
 
+
         switch (key)
         {
             case "test1":

Src/FastData.Cli.Tests/CommandOutputs/rust_-k UInt8_Files_Integers.input.verified.txt

@@ -17,6 +17,7 @@ impl MyData {
             return false;
         }
 
+
         if key == 42 || key == 62 || key == 80 || key == 100 {
             return true;
         }

Src/FastData.Cli.Tests/CommandOutputs/rust_-s HashTable_Files_Strings.input.verified.txt

@@ -46,6 +46,7 @@ const ENTRIES: [e; 2] = [
             return false;
         }
 
+
         let hash = unsafe { Self::get_hash(key) };
         let index = (hash % 2) as usize;
 

Src/FastData.Cli.Tests/CommandOutputs/rust_Files_Strings.input.verified.txt

@@ -17,6 +17,7 @@ impl MyData {
             return false;
         }
 
+
         if key == "test1" || key == "test2" {
             return true;
         }