Brudixy

High-performance in-memory data tables for .NET, with optional compile-time code generation from YAML schemas.

Brudixy is aimed at two kinds of consumers:

1. Runtime users: you want a fast in-memory table with indexing, querying, relations, expressions, and serialization.
2. Schema-driven users: you want strongly-typed tables/datasets generated from YAML at build time.

---

Highlights

• Fast in-memory tables with indexes (single + multi-column)
• Fluent query API (table.Rows.Where(...).Equals(...))
• Computed columns / filters via expressions
• Relations with cascade rules
• Change tracking, transactions, and optional change logging
• JSON/XML serialization
• Extensible metadata: XProperties + row/cell annotations

---

Comparison with System.Data DataSet/DataTable

Brudixy is designed for fast in-memory processing and schema-driven code generation. It is not a drop-in replacement for ADO.NET tables, but it covers many of the same data-shaping scenarios.

Advantages / where Brudixy shines

• Dataset-like composition at the table level: a DataTable can own child tables via AddTable(...) / GetTable(...) and manage relations by name.
• Code generation from YAML: compile-time, strongly-typed tables/datasets without the legacy designer workflow.
• Index-first design: built-in single and multi-column indexes optimized for fast lookups.
• Expressions + computed columns: expression engine for filters and derived values.
• Rich metadata: table/row/column XProperties and row/cell annotations for UI, validation, and auditing.
• Lightweight, in-memory focus: optimized for high-performance runtime operations rather than database plumbing.

Disadvantages / missing features vs System.Data

• Not a drop-in ADO.NET replacement: APIs like DataAdapter, DataView, BindingSource, and typed DataSet designers are not part of Brudixy.
• Fewer ecosystem integrations: System.Data has decades of tooling support across ORMs, designers, and data-binding stacks.
• Different mental model: Brudixy exposes a composable DataTable (table-as-dataset), which may require adaptation if you rely on DataSet-specific patterns.

---

Dependencies:

• Akade.IndexedSet
• Konsarpoo

---

Packages

Runtime:

• Brudixy – main consumer package (brings Brudixy.Core + Brudixy.Interfaces)

Build-time (source generators):

• Brudixy.TypeGenerator – YAML-driven generator that produces strongly-typed DataSets/DataTables

---

Install

Runtime only

Add the Brudixy package.

Runtime + code generation

Add:

• Brudixy (runtime)
• Brudixy.TypeGenerator as an analyzer (PrivateAssets="all")
• YAML schema files as AdditionalFiles

---

Quickstart (runtime)

1) Create a table + add columns

using Brudixy;

var table = new DataTable("Users");

table.AddColumn("Id", TableStorageType.Int32, unique: true);
table.AddColumn("Name", TableStorageType.String);

table.SetPrimaryKeyColumn("Id");

2) Add a row

var r = table.NewRow();
r["Id"] = 1;
r["Name"] = "Alice";

table.AddRow(r);

3) Query

var row = table.GetRow("Id", 1);
var name = row.Field<string>("Name");

var filtered = table.Rows
    .Where("Name").AsString().StartsWith("Al")
    .ToData();

---

Quickstart (YAML schema → generated types)

1. Put schema files under Schemas/.
2. Include them as AdditionalFiles.
3. Add the generator as an analyzer.

Example .csproj

<ItemGroup>
  <PackageReference Include="Brudixy" Version="1.0.0" />

  <!-- build-time only -->
  <PackageReference Include="Brudixy.TypeGenerator" Version="1.0.0" PrivateAssets="all" />

  <AdditionalFiles Include="Schemas\**\*.brudixy.yaml" />
</ItemGroup>

Single-table schema (*.st.brudixy.yaml)

---
Table: Users
PrimaryKey:
  - Id
Columns:
  Id: Int32
  Name: String

Generator output becomes available during compilation. You don’t ship the generator at runtime.

---

Core concepts

DataTable, DataRow, and DataRowContainer

• DataTable is the main structure.
• DataRow is a row stored in a table.
• DataRowContainer is a detached, serializable/editable row container (useful for JSON/XML round-trips, UI, and patching).

Important: Brudixy tables are composable.

• A DataTable can act as a dataset container (a collection of child tables). This is why you’ll see APIs like AddTable(...), GetTable(...), and relations referencing table names.
• A DataTable can also be stored as a value:
  • inside a cell (column value)
  • inside XProperties (table/row/column XProperties and XProperty annotations)

This makes it possible to model nested documents/graphs directly in-memory.

Indexes

Indexes are the key to performance.

• Single-column: AddIndex("Email") / AddIndex("Email", unique: true)
• Multi-column: AddMultiColumnIndex(new[] { "A", "B" }, unique: false)

Arrays and ranges

Columns can store:

• a single value (Simple)
• arrays (Array)
• ranges (Range)

In YAML you can specify this via TypeModifier: Array|Range.

---

Expressions (computed columns + filters)

Computed columns

table.AddColumn("FullName", TableStorageType.String, dataExpression: "FirstName + ' ' + LastName");

Filtering

var rows = table.Select("Id = 5 AND Len(Name) > 2");

Check a filter against a single row

bool ok = table.Rows.First().CheckFilter("Id = 5 AND Name <> ''");

Register custom functions

You can extend the expression engine:

using Brudixy.Expressions;

FunctionRegistry.Registry.RegisterFunction("IsEven", _ => new IsEvenFunction());

---

Relations (including cascade rules)

Brudixy supports relations between tables (dataset style) and self-relations.

var ds = new DataTable("MyDs");
var parent = ds.AddTable("Parent");
var child = ds.AddTable("Child");

parent.AddColumn("Id", TableStorageType.Int32, unique: true);
child.AddColumn("ParentId", TableStorageType.Int32);

ds.AddRelation(
    relationName: "FK_Child_Parent",
    parentKey: ("Parent", "Id"),
    childKey:  ("Child", "ParentId"),
    relationType: RelationType.OneToMany,
    constraintUpdate: Rule.Cascade,
    constraintDelete: Rule.Cascade,
    acceptRejectRule: AcceptRejectRule.Cascade);

Notes:

• constraintUpdate controls parent key update behavior.
• constraintDelete controls delete cascades.
• acceptRejectRule controls AcceptChanges() / RejectChanges() propagation.

---

Change tracking, transactions, and logging

Transactions

var tran = table.StartTransaction();

// ... make changes

tran.Rollback(); // or tran.Commit();

Change logging (audit stream)

using var _ = table.StartLoggingChanges("Import #42");
// ... change rows / xprops
var log = table.GetLoggedChanges();

Transaction connection:

• Each log entry can carry a TranId.
• On rollback, Brudixy removes log entries for rolled-back transactions.

---

Defaults, nullability, and safe conversion

• row["Col"] returns the raw stored value (can be null).
• row.Field<T>("Col") returns a typed value and can apply safe conversions.
• For nullable strings/arrays, Field<string> can return "" and Field<int[]> can return an empty array even when stored value is null.

Arrays are treated as immutable values:

• assignment copies the array
• FieldArray<T>() provides a cached reference for fast repeated access

---

Debugging (visualizers)

Brudixy is debugger-friendly:

• DataTable has a helpful DebuggerDisplay with counts/index info.
• DataRow / DataRowContainer have a debug view (DataRowDebugView) that shows:
  • column values
  • ages / changed fields
  • annotations and XProperties
  • parent/child summaries (when relations exist)

---

Comparing rows and containers

Recommended pattern when validating serialization or container logic:

var row = table.GetRowByPk(new (1, 1));
var container = row.ToContainer();

var cmp = DataRowContainer.CompareDataRows(row, container);
if (cmp.cmp != 0)
{
    throw new Exception(cmp.ToString());
}

---

YAML schemas (runtime loading)

If you want runtime-loaded schemas (plugins/config-driven), use:

• LoadSchemaFromYaml(string yaml)
• LoadSchemaFromYamlFile(string path)
• ToYaml()

Brudixy supports a compact schema form:

Table: SimpleTable
Columns:
  Id: Int32
  Name: String
PrimaryKey:
  - Id

For advanced column options (default values, max length, expressions, etc.), use the verbose form.

---

Dapper support (vendored)

Brudixy includes a built-in copy of Dapper (SqlMapper) in the Brudixy assembly.

using System.Data;
using Brudixy;

using var conn = /* IDbConnection */;
var rows = conn.Query<MyPoco>("select Id, Name from Users where Id = @id", new { id = 1 });

If you also reference the external Dapper package, you may get ambiguous extension method resolution.

---

For contributors / maintainers

• Build and pack instructions live in PROJECT_SETUP.md and NUGET_PUBLISHING.md.
• This README is intended for package consumers.