The Record Definition Input transform (plugin id RecordDefinitionDataInput) reads actual data rows from a Hop data catalog record definition. It uses the same source pipeline builders as Preview data in the Data Catalog perspective (database table, CSV, Parquet, Iceberg, composite .hsm query, or JSON extraction).

It is an input transform (category Input) and appears in the transform palette with the data catalog icon.

Note
 The older transform that streams catalog metadata (definition names, field layouts for DDL, etc.) is now named Get Record Definition Names (plugin id RecordDefinitionInput).

When to use it

  • Build a source-side pipeline that starts from a catalog-published feed without hard-coding connection/schema/table details.

  • Seed pipelines generated from a source-model pipeline source (see Source modeler overview).

  • Sample or load source data for custom transforms (filter, convert, aggregate) before publishing the result as a pipeline-backed vault feed.

Prerequisites

  • A configured data catalog connection.

  • A record definition that supports preview (physical location or composite/JSON pointers), typically a DV_SOURCE.

Dialog options

Option Description

Transform name

Name of this transform in the pipeline graph.

Catalog connection

Required. Catalog connection used to look up the record definition.

Select from previous transform input?

When disabled (default), the transform reads one fixed definition (namespace + name values). When enabled, each input row supplies namespace and name fields and the transform streams that definition’s data before moving to the next input row.

Namespace field / Name field

Enabled when selecting from input. Incoming field names that hold the catalog namespace and definition name.

Namespace value / Name value

Enabled when not selecting from input. Fixed catalog key for the definition to read.

Row limit (0 = all)

Optional limit passed into the generated source input (for example Table Input row limit). Use 0 for a full read.

Output fields

Output columns follow the field layout stored on the catalog record definition (names, Hop types, length, precision). At runtime, source columns are mapped onto that layout by name.

When Select from previous transform input? is enabled, the design-time field list may be empty until the pipeline runs (each input row can point at a different definition).