Skip to main content

​
Overview

Automap suggests how your source data should map to a data type. Rather than finding each table and dragging each field by hand, you run Automap and get a ranked set of suggestions. It covers the primary source, field mappings, joins to related sources, and source filters. You review the suggestions, accept the ones you want, and save. You can run Automap in two places:

In the Mapping Editor

Run Automap on one data type from the Automap button in the mapping editor

Across a Pipeline

Step through every unmapped silver data type in a pipeline, in dependency order
Required permissions
  • Data Pipelines: Access
  • Manage Silver Entities: Update
  • Manage Production Pipelines (for production pipelines and the main pipeline) or Manage Dev Pipelines (for development pipelines): Update
These are the same permissions you need to map fields by hand. See roles and permissions.

​
What Automap Suggests

Automap suggests mappings only for active fields that aren’t mapped yet. Your existing mappings stay as they are.

​
How Suggestions Are Rated

Every suggestion has a confidence level. Automap bases it on several independent signals:
  • Schema expectations: Entegrata’s curated knowledge of which source tables and columns populate each standard data type field
  • Historical mappings: mapping patterns from pipelines that are already deployed
  • Name similarity: how closely a column name matches the field name
  • Usage frequency: how often a table is used as a source
  • Column statistics: profiles of the data in each column, such as null rate and uniqueness
When more signals agree on a suggestion, its confidence goes up. Suggestions backed by a schema expectation rank highest. Automap leaves out columns it can’t use, such as columns with an incompatible data type and columns that are always empty or always hold the same value. Expand the provider count on a suggestion (for example, 3 providers) to see which signals contributed.

​
Confidence Bands

Each suggestion is labeled with a confidence band: Suggestions backed by a schema expectation show the Entegrata logo and the label Schema expectation. Table filter suggestions show a funnel icon.

​
What Gets Applied Automatically

Automap applies its strongest suggestions for you:
  • high-confidence schema expectation suggestions
  • other high-confidence suggestions whose source is already on the canvas, or can be reached through a high-confidence join
Everything else waits for you to review it.
Applied suggestions are unsaved edits on the canvas, just like edits you make by hand. Nothing reaches your pipeline until you click Save, and you can undo any of it first. See Saving Your Work.

​
Limitations

  • Automap works on silver data types.
  • Automap suggests mappings from tables already in your source catalog. If a table hasn’t been collected yet, Automap can’t suggest it. See enabling resources.

Automap in the Mapping Editor

Run Automap on a single data type and review its suggestions

Automap a Pipeline

Walk through every unmapped silver data type in a pipeline

Mapping Fields

Map fields by hand

Configuring Sources

Primary sources, related sources, and joins