Open.Text

A set of useful extensions for working with strings, string-segments, spans, enums, and value formatting.

Features

• Avoids allocation wherever possible.
• v3.x is a major overhaul with much improved methods and expanded tests and coverage.
• v4.x favored use of Microsoft.Extensions.Primitives.StringSegments for string manipulation.
• NEW: Zero-allocation *NoAlloc methods available in the separate Open.Text.ZLinq package!
• NEW: Roslyn analyzers to detect inefficient string patterns and suggest modern alternatives!

---

📊 Benchmark Summary

Comparing string.Split() (BCL) vs SplitAsSegments (IEnumerable) vs SplitAsSegmentsNoAlloc (ValueEnumerable via ZLinq):

Category	Method	Time	Allocated
Count	BCL Split + LINQ Count	46.9 ns	256 B
	SplitAsSegments + LINQ Count	61.9 ns	88 B
	SplitAsSegmentsNoAlloc + ZLinq Count	55.3 ns	0 B ✅
LINQ-Chain	SplitAsSegmentsNoAlloc + ZLinq	50.8 ns	0 B ✅
	BCL + System.Linq	65.5 ns	304 B
	SplitAsSegments + System.Linq	104.4 ns	152 B
Large-Foreach	BCL Split (1000 items)	7,342 ns	47,952 B
	SplitAsSegments (1000 items)	11,740 ns	88 B
	SplitAsSegmentsNoAlloc (1000 items)	11,557 ns	0 B ✅
Small-Foreach	BCL Split	42.5 ns	256 B
	SplitAsSegmentsNoAlloc	45.6 ns	0 B ✅
	SplitAsSegments	72.4 ns	88 B
Seq-Split	SplitAsSegmentsNoAlloc(string)	177.6 ns	0 B ✅
	SplitAsSegments(string)	215.2 ns	128 B
	BCL Split(string)	227.1 ns	696 B

Key Takeaway: The SplitAsSegmentsNoAlloc methods achieve zero heap allocations when iterating StringSegment values—ideal for high-throughput scenarios where GC pressure matters.

Note: Regex-based split methods have unavoidable Match object allocations.

---

Open.Text.ZLinq - Zero-Allocation Extension Package

For scenarios requiring true zero-allocation string operations, install the companion package:

dotnet add package Open.Text.ZLinq

This package provides *NoAlloc extension methods that return ZLinq ValueEnumerable<TEnumerator, T> structs instead of heap-allocated enumerables. These methods integrate seamlessly with ZLinq for zero-allocation LINQ operations.

Available Methods

• SplitAsSegmentsNoAlloc(char) - Split by character
• SplitAsSegmentsNoAlloc(string) - Split by string sequence
• SplitAsSegmentsNoAlloc(Regex) - Split by regex pattern
• JoinNoAlloc(...) - Join segments with separator
• ReplaceNoAlloc(...) / ReplaceAsSegmentsNoAlloc(...) - Replace sequences
• AsSegmentsNoAlloc(Regex) - Get regex matches as segments

Example

using Open.Text; using ZLinq; // Zero-allocation split and filter var count = "a,b,c,d,e" .SplitAsSegmentsNoAlloc(',') .Where(s => s.Length > 0) .Count(); // No heap allocations!

---

🔍 Roslyn Analyzers

Open.Text now includes Roslyn analyzers that help you write more efficient code by detecting common string manipulation anti-patterns and suggesting better alternatives using spans and string segments.

Installation

dotnet add package Open.Text.Analyzers

What it does

The analyzers detect patterns like:

• .Substring() → suggests .AsSpan() or span slicing
• .Split() → suggests .SplitAsSegments() or .SplitToEnumerable() to reduce allocations
• .Split()[0] → suggests .FirstSplit() to avoid array allocation
• String concatenation in loops → suggests StringBuilder
• .Trim().Equals() → suggests .TrimEquals() to avoid intermediate string

See the Analyzers README for complete documentation.

---

Regex Extensions

ReadOnlySpan<char> Capture.AsSpan() Enumerable<StringSegment> Regex.AsSegments(string input) string GroupCollection.GetValue(string groupName) ReadOnlySpan<char> GroupCollection.GetValueSpan(string groupName) IEnumerable<StringSegment> string.Split(Regex pattern)

---

String vs Span Equality

Optimized .Equals(...) extension methods for comparing spans and strings.

---

String & Span Splitting

SplitToEnumerable

Returns each string segment of the split through an enumerable instead of all at once in an array.

SplitAsMemory

Produces an enumerable where each segment is yielded as a ReadOnlyMemory<char>.

SplitAsSegment

Produces an enumerable where each segment is yielded as a StringSegment.

---

Trimming

TrimStartPattern & TrimEndPattern

Similar to their character trimming counterparts, these methods can trim sequences of characters or regular expression patterns.

---

StringBuilder Extensions

Extensions for:

• adding segments with separators.
• adding spans without creating a string first.
• converting enumerables to a StringBuilder.

---

StringSegment Extensions

Extensions for:

• .Trim(char) and .Trim(ReadOnlySpan<char>).
• finding and pivoting from sequences without allocation.

---

StringComparable & SpanComparable Extensions

if(myString.AsCaseInsensitive()=="HELLO!") { }

instead of

if(myString.Equals("HELLO!", StringComparison.OrdinalIgnoreCase)) { }

---

EnumValue<TEnum> & EnumValueIgnoreCase<TEnum>

Implicit conversion makes it easy. Optimized methods make it fast.

Consider the following:

enum Greek { Alpha, Beta, Gamma } void DoSomethingWithGreek(Greek value) { } DoSomethingWithGreek(Greek.Alpha);

It's nice that Greek is an enum because it won't be null, and it has to be one of the values. But what if you want to write a single function that will take an Greek or a string? This gets problematic as the string value has to be parsed and you'll likely need an overload.

EnumValue<TEnum> solves this problem:

enum Greek { Alpha, Beta, Gamma } void DoSomethingWithGreek(EnumValue<Greek> value) { } // Both work fine. DoSomethingWithGreek("Alpha"); DoSomethingWithGreek(Greek.Alpha); // Throws an ArgumentException: DoSomethingWithGreek("Theta");

The implicit conversion between a string and EnumValue<TEnum> make this possible.

If you need to allow for case-insensitive comparison then simply use EnumValueCaseIgnored<TEnum> instead.

The performance is outstanding as it uses the length of the names to build a tree in order to parse values and uses an expression tree instead of calling .ToString() on the value.

---

And more ...

string.Supplant(...)

An alternative to String.Format that takes an array of values.

string.ReplaceWhiteSpace(...)

A shortcut for replacing whitespace with a Regex.

string.ToMetricString(...)

Returns an abbreviated metric representation of a number.

ToByteString(...)

Returns an abbreviated metric representation of a quantity of bytes.

ToPercentString(...)

Shortcut for formating to a percent.

ToNullIfWhiteSpace()

Allows for simple null operators if a string is empty or whitespace.