Redis
redis-vl-dotnet Documentation redis-vl-dotnet

redis-vl-dotnet

YAML Schema Loading

Use YAML when you want a schema definition that can be checked into the repository, reused by examples, and loaded by the CLI or application code without duplicating constructor logic.

Loading APIs

SearchSchema exposes two entry points:

  • SearchSchema.FromYaml(string yaml)

  • SearchSchema.FromYamlFile(string path)

The JSON storage example uses FromYamlFile:

var schemaPath = Path.Combine(AppContext.BaseDirectory, "schema.yaml");
var schema = SearchSchema.FromYamlFile(schemaPath);

YAML shape

A schema file has index metadata plus a fields collection:

index:
  name: json-storage-example-idx
  prefixes:
    - "json-example:"
    - "json-example-archive:"
  key_separator: "|"
  storage_type: json
  stopwords: []
  max_text_fields: true
  no_offsets: true
  no_highlight: true
  skip_initial_scan: true
fields:
  - name: title
    type: text
    sortable: true
    weight: 2.0
  - name: genre
    type: tag
  - name: year
    type: numeric
    sortable: true

Supported index keys include:

  • name

  • prefix or prefixes

  • key_separator

  • storage_type

  • stopwords

  • max_text_fields

  • temporary

  • no_offsets

  • no_highlight

  • no_fields

  • no_frequencies

  • skip_initial_scan

Supported field type values:

  • text

  • tag

  • numeric

  • geo

  • vector

Vector fields must include an attrs block with algorithm, data type, dimensions, and distance metric.

Validation behavior

The loader rejects invalid YAML or unsupported values early with ArgumentException.

Common validation failures:

  • missing index or fields

  • defining both index.prefix and index.prefixes

  • unsupported field or enum values

  • vector definitions without attrs

  • invalid combinations such as HNSW-only settings on a FLAT vector field

The repository tests under tests/RedisVL.Tests/Schema cover both accepted and rejected YAML shapes. The fixture advanced-schema.yaml is the best reference when you need every supported index-level toggle in one file.

When to prefer YAML

YAML is the better fit when:

  • the schema should be visible without recompiling code

  • examples, docs, and CLI commands should point to the same schema artifact

  • you want to keep index metadata in a declarative file rather than in application startup code

Use constructor-based schema definitions when the field list is generated dynamically or is tightly coupled to application code.

Related pages