Implementation:Datahub project Datahub ProtobufExtensionFieldVisitor

Knowledge Sources	Datahub_project_Datahub
Domains	Protobuf_Integration, Schema_Field_Metadata
Last Updated	2026-02-10 00:00 GMT

Overview

Description

ProtobufExtensionFieldVisitor extends SchemaFieldVisitor to produce DataHub SchemaField objects enriched with metadata extracted from protobuf extension options. For each protobuf field, it:

1. Detects primary keys -- Checks field options for any descriptor name matching the pattern primary_key (case-insensitive).
2. Extracts tags -- Collects TagAssociation objects from both field-level options and promoted message-level options (from nested message types referenced by the field).
3. Extracts glossary terms -- Collects GlossaryTermAssociation objects similarly from field and promoted message options.
4. Builds SchemaField -- Constructs the SchemaField with field path, nullability (inverse of primary key), description (including enum value documentation), native data type, global tags, and glossary terms.
5. Handles enum descriptions -- For enum-typed fields, appends a formatted listing of enum values and their comments to the field description.

The visitor processes all graph paths to each field, producing a pair of (SchemaField, sortOrder) for each path.

Usage

Used as the primary field visitor during protobuf schema conversion to generate the SchemaField array for the schemaMetadata aspect of a DataHub dataset.

Code Reference

Source Location

metadata-integration/java/datahub-protobuf/src/main/java/datahub/protobuf/visitors/field/ProtobufExtensionFieldVisitor.java

Signature

public class ProtobufExtensionFieldVisitor extends SchemaFieldVisitor {

    @Override
    public Stream<Pair<SchemaField, Double>> visitField(
        ProtobufField field, VisitContext context)
}

Import

import datahub.protobuf.visitors.field.ProtobufExtensionFieldVisitor;

I/O Contract

Inputs

Parameter	Type	Description
field	ProtobufField	The protobuf field being visited
context	VisitContext	Visit context providing the graph, registry, audit stamp, and path computation

Outputs

Return Type	Description
Stream<Pair<SchemaField, Double>>	Pairs of constructed SchemaField objects and their computed sort order values (one pair per graph path to the field)

Each SchemaField contains:

• fieldPath -- computed from the graph path
• nullable -- true unless the field is a primary key
• isPartOfKey -- true if primary key option is present
• description -- field comment with appended enum values if applicable
• nativeDataType -- the protobuf native type string
• type -- the DataHub schema field data type
• globalTags -- tags from field options and promoted message options
• glossaryTerms -- terms from field options and promoted message options

Usage Examples

ProtobufExtensionFieldVisitor fieldVisitor = new ProtobufExtensionFieldVisitor();

// Used within ProtobufGraph traversal
Stream<Pair<SchemaField, Double>> fieldResults =
    fieldVisitor.visitField(protobufField, visitContext);

Related Pages

• Datahub_project_Datahub_ProtobufExtensionUtil -- Provides tag and term extraction from options
• Datahub_project_Datahub_ProtobufDescriptorUtils -- Provides getFieldOptions and getMessageOptions
• Datahub_project_Datahub_ProtobufDatasetVisitor -- Top-level visitor that composes field visitors
• Datahub_project_Datahub_TagVisitor -- Related visitor for dataset-level tag creation