Skip to content

Commit 545c7d4

Browse files
GenboxGenbox
committed
Add an indicator for HashSets that store hash code, that when using identity hashing, we don't need to store the hash code
1 parent 938205b commit 545c7d4

File tree

8 files changed

+30
-17
lines changed

8 files changed

+30
-17
lines changed

‎Src/FastData.InternalShared/Helpers/TestHelper.cs‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -99,9 +99,9 @@ public static GeneratorSpec Generate<T>(Func<string, ICodeGenerator> func, TestV
9999
if (vector.Type == typeof(EytzingerSearchStructure<>))
100100
return Generate(func, props, vector, new EytzingerSearchStructure<T>(props.DataType, StringComparison.Ordinal));
101101
if (vector.Type == typeof(HashSetChainStructure<>))
102-
return Generate(func, props, vector, new HashSetChainStructure<T>(HashData.Create(data, props.DataType, 1)));
102+
return Generate(func, props, vector, new HashSetChainStructure<T>(HashData.Create(data, props.DataType, 1), props.DataType));
103103
if (vector.Type == typeof(HashSetPerfectStructure<>))
104-
return Generate(func, props, vector, new HashSetPerfectStructure<T>(HashData.Create(data, props.DataType, 1)));
104+
return Generate(func, props, vector, new HashSetPerfectStructure<T>(HashData.Create(data, props.DataType, 1), props.DataType));
105105
if (vector.Type == typeof(HashSetLinearStructure<>))
106106
return Generate(func, props, vector, new HashSetLinearStructure<T>(HashData.Create(data, props.DataType, 1)));
107107
if (vector.Type == typeof(KeyLengthStructure<>))

‎Src/FastData.Testbed/Tests/GPerfTest.cs‎

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -43,10 +43,11 @@ public static void ProduceOutputs(string path)
4343
StringProperties props = DataAnalyzer.GetStringProperties(data);
4444

4545
GPerfAnalyzer analyzer = new GPerfAnalyzer(data.Length, props, new GPerfAnalyzerConfig(), new Simulator(new SimulatorConfig()), factory.CreateLogger<GPerfAnalyzer>());
46+
4647
// Candidate hashFunc = analyzer.GetCandidates(data).First(); //TODO: use
4748

4849
HashData hashData = HashData.Create(data, DataType.String, 1);
49-
HashSetPerfectStructure<string> structure = new HashSetPerfectStructure<string>(hashData);
50+
HashSetPerfectStructure<string> structure = new HashSetPerfectStructure<string>(hashData, DataType.String);
5051
structure.Create(ref data);
5152
}
5253
}

‎Src/FastData/FastDataGenerator.cs‎

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -130,9 +130,9 @@ public static string Generate<T>(ReadOnlySpan<T> data, FastDataConfig fdCfg, ICo
130130
HashData hashData = HashData.Create(data, props.DataType, fdCfg.HashCapacityFactor);
131131

132132
if (hashData.HashCodesPerfect)
133-
return Generate(generator, genCfg, new HashSetPerfectStructure<T>(hashData), data);
133+
return Generate(generator, genCfg, new HashSetPerfectStructure<T>(hashData, props.DataType), data);
134134

135-
return Generate(generator, genCfg, new HashSetChainStructure<T>(hashData), data);
135+
return Generate(generator, genCfg, new HashSetChainStructure<T>(hashData, props.DataType), data);
136136
}
137137
case StructureType.Array:
138138
return Generate(generator, genCfg, new ArrayStructure<T>(), data);
@@ -145,9 +145,9 @@ public static string Generate<T>(ReadOnlySpan<T> data, FastDataConfig fdCfg, ICo
145145
HashData hashData = HashData.Create(data, props.DataType, fdCfg.HashCapacityFactor);
146146

147147
if (hashData.HashCodesPerfect)
148-
return Generate(generator, genCfg, new HashSetPerfectStructure<T>(hashData), data);
148+
return Generate(generator, genCfg, new HashSetPerfectStructure<T>(hashData, props.DataType), data);
149149

150-
return Generate(generator, genCfg, new HashSetChainStructure<T>(hashData), data);
150+
return Generate(generator, genCfg, new HashSetChainStructure<T>(hashData, props.DataType), data);
151151
}
152152
default:
153153
throw new InvalidOperationException($"Unsupported DataStructure {fdCfg.StructureType}");

‎Src/FastData/Generators/Contexts/HashSetChainContext.cs‎

Lines changed: 5 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -7,11 +7,15 @@ namespace Genbox.FastData.Generators.Contexts;
77
/// <typeparam name="T">The type of elements in the data array.</typeparam>
88
/// <param name="buckets">The array of bucket indices.</param>
99
/// <param name="entries">The array of hash set entries.</param>
10-
public sealed class HashSetChainContext<T>(int[] buckets, HashSetEntry<T>[] entries) : IContext<T>
10+
/// <param name="storeHashCode">If set to true, you should only generate a hash set that checks the value.</param>
11+
public sealed class HashSetChainContext<T>(int[] buckets, HashSetEntry<T>[] entries, bool storeHashCode) : IContext<T>
1112
{
1213
/// <summary>Gets the array of bucket indices.</summary>
1314
public int[] Buckets { get; } = buckets;
1415

1516
/// <summary>Gets the array of hash set entries.</summary>
1617
public HashSetEntry<T>[] Entries { get; } = entries;
18+
19+
/// <summary>Indicates whether the hash set should store the hash code or only the value.</summary>
20+
public bool StoreHashCode { get; } = storeHashCode;
1721
}

‎Src/FastData/Generators/Contexts/HashSetPerfectContext.cs‎

Lines changed: 5 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -5,8 +5,12 @@ namespace Genbox.FastData.Generators.Contexts;
55
/// <summary>Provides a context for perfect hash set-based data structures.</summary>
66
/// <typeparam name="T">The type of keys in the key-value pairs.</typeparam>
77
/// <param name="data">The array of key-value pairs and their hash codes.</param>
8-
public sealed class HashSetPerfectContext<T>(KeyValuePair<T, ulong>[] data) : IContext<T>
8+
/// <param name="storeHashCode">If set to true, you should only generate a hash set that checks the value.</param>
9+
public sealed class HashSetPerfectContext<T>(KeyValuePair<T, ulong>[] data, bool storeHashCode) : IContext<T>
910
{
1011
/// <summary>Gets the array of items and their hash codes.</summary>
1112
public KeyValuePair<T, ulong>[] Data { get; } = data;
13+
14+
/// <summary>Indicates whether the hash set should store the hash code or only the value.</summary>
15+
public bool StoreHashCode { get; } = storeHashCode;
1216
}

‎Src/FastData/Internal/Structures/HashSetChainStructure.cs‎

Lines changed: 4 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,11 +1,13 @@
1+
using Genbox.FastData.Enums;
12
using Genbox.FastData.Generators.Contexts;
23
using Genbox.FastData.Generators.Contexts.Misc;
4+
using Genbox.FastData.Generators.Extensions;
35
using Genbox.FastData.Internal.Abstracts;
46
using Genbox.FastData.Internal.Misc;
57

68
namespace Genbox.FastData.Internal.Structures;
79

8-
internal sealed class HashSetChainStructure<T>(HashData hashData) : IStructure<T, HashSetChainContext<T>>
10+
internal sealed class HashSetChainStructure<T>(HashData hashData, DataType dataType) : IStructure<T, HashSetChainContext<T>>
911
{
1012
public HashSetChainContext<T> Create(ref ReadOnlySpan<T> data)
1113
{
@@ -27,6 +29,6 @@ public HashSetChainContext<T> Create(ref ReadOnlySpan<T> data)
2729
bucket = i + 1;
2830
}
2931

30-
return new HashSetChainContext<T>(buckets, entries);
32+
return new HashSetChainContext<T>(buckets, entries, !dataType.IsIdentityHash());
3133
}
3234
}

‎Src/FastData/Internal/Structures/HashSetPerfectStructure.cs‎

Lines changed: 4 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1,10 +1,12 @@
1+
using Genbox.FastData.Enums;
12
using Genbox.FastData.Generators.Contexts;
3+
using Genbox.FastData.Generators.Extensions;
24
using Genbox.FastData.Internal.Abstracts;
35
using Genbox.FastData.Internal.Misc;
46

57
namespace Genbox.FastData.Internal.Structures;
68

7-
internal sealed class HashSetPerfectStructure<T>(HashData hashData) : IStructure<T, HashSetPerfectContext<T>>
9+
internal sealed class HashSetPerfectStructure<T>(HashData hashData, DataType dataType) : IStructure<T, HashSetPerfectContext<T>>
810
{
911
public HashSetPerfectContext<T> Create(ref ReadOnlySpan<T> data)
1012
{
@@ -20,6 +22,6 @@ public HashSetPerfectContext<T> Create(ref ReadOnlySpan<T> data)
2022
for (int i = 0; i < data.Length; i++)
2123
pairs[hashCodes[i] % size] = new KeyValuePair<T, ulong>(data[i], hashCodes[i]);
2224

23-
return new HashSetPerfectContext<T>(pairs);
25+
return new HashSetPerfectContext<T>(pairs, !dataType.IsIdentityHash());
2426
}
2527
}

‎Src/FastData/PublicAPI.Unshipped.txt‎

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -60,8 +60,8 @@ Genbox.FastData.Generators.Contexts.EytzingerSearchContext<T>.EytzingerSearchCon
6060
Genbox.FastData.Generators.Contexts.HashSetChainContext<T>
6161
Genbox.FastData.Generators.Contexts.HashSetChainContext<T>.Buckets.get -> int[]!
6262
Genbox.FastData.Generators.Contexts.HashSetChainContext<T>.Entries.get -> Genbox.FastData.Generators.Contexts.Misc.HashSetEntry<T>[]!
63-
Genbox.FastData.Generators.Contexts.HashSetChainContext<T>.HashSetChainContext() -> void
64-
Genbox.FastData.Generators.Contexts.HashSetChainContext<T>.HashSetChainContext(int[]! buckets, Genbox.FastData.Generators.Contexts.Misc.HashSetEntry<T>[]! entries) -> void
63+
Genbox.FastData.Generators.Contexts.HashSetChainContext<T>.HashSetChainContext(int[]! buckets, Genbox.FastData.Generators.Contexts.Misc.HashSetEntry<T>[]! entries, bool storeHashCode) -> void
64+
Genbox.FastData.Generators.Contexts.HashSetChainContext<T>.StoreHashCode.get -> bool
6565
Genbox.FastData.Generators.Contexts.HashSetLinearContext<T>
6666
Genbox.FastData.Generators.Contexts.HashSetLinearContext<T>.Buckets.get -> Genbox.FastData.Generators.Contexts.Misc.HashSetBucket[]!
6767
Genbox.FastData.Generators.Contexts.HashSetLinearContext<T>.Data.get -> T[]!
@@ -70,8 +70,8 @@ Genbox.FastData.Generators.Contexts.HashSetLinearContext<T>.HashSetLinearContext
7070
Genbox.FastData.Generators.Contexts.HashSetLinearContext<T>.HashSetLinearContext(T[]! data, Genbox.FastData.Generators.Contexts.Misc.HashSetBucket[]! buckets, ulong[]! hashCodes) -> void
7171
Genbox.FastData.Generators.Contexts.HashSetPerfectContext<T>
7272
Genbox.FastData.Generators.Contexts.HashSetPerfectContext<T>.Data.get -> System.Collections.Generic.KeyValuePair<T, ulong>[]!
73-
Genbox.FastData.Generators.Contexts.HashSetPerfectContext<T>.HashSetPerfectContext() -> void
74-
Genbox.FastData.Generators.Contexts.HashSetPerfectContext<T>.HashSetPerfectContext(System.Collections.Generic.KeyValuePair<T, ulong>[]! data) -> void
73+
Genbox.FastData.Generators.Contexts.HashSetPerfectContext<T>.HashSetPerfectContext(System.Collections.Generic.KeyValuePair<T, ulong>[]! data, bool storeHashCode) -> void
74+
Genbox.FastData.Generators.Contexts.HashSetPerfectContext<T>.StoreHashCode.get -> bool
7575
Genbox.FastData.Generators.Contexts.KeyLengthContext<T>
7676
Genbox.FastData.Generators.Contexts.KeyLengthContext<T>.KeyLengthContext() -> void
7777
Genbox.FastData.Generators.Contexts.KeyLengthContext<T>.KeyLengthContext(System.Collections.Generic.List<string!>?[]! lengths, bool lengthsAreUniq, uint minLength, uint maxLength) -> void

0 commit comments

Comments
 (0)