Pipelines

Map source columns to target columns

The mapping editor is where you shape the destination schema. Every change saves as you make it. There is no separate Save button.

In the app: Data Model › Pipelines › your Pipeline

[screenshot]
The mapping editor with one table expanded, showing its column grid and an AI suggestion with Accept and Reject buttons.

Open the mapping editor

On the Pipelines page, click a Pipeline’s name or its View table mappings button. The header shows the Pipeline name and a count such as Mapping Editor — 12 tables, 148 columns.

The toolbar

Button What it does Page
Sync from Source Picks up tables, views, and columns added to or removed from the sources. Keep a Pipeline in sync
Re-run Reopens the create dialog so you can change Pipeline-wide settings. Re-run a Pipeline
Validate Checks the mappings and lists errors and warnings. Validate a Pipeline
Finalize Builds the destination schema. Enabled after a validation with no errors. Finalize a Pipeline
Re-finalize Appears only on a finalized Pipeline that has drifted or has out-of-date load procedures. Keep a Pipeline in sync

The editor has three tabs: Table Mappings, Documentation, and Summary.

Table Mappings tab

One row per source table or view. Use Search tables... to filter by source schema, source table, or destination table name.

Column What it does
Include checkbox Clear it to leave the table out. Excluded rows are dimmed and show (excluded).
Source Schema, Type, Source Table Where the row comes from. Type is Table or View. A Stale tag means the source no longer exists.
SK Adds a generated surrogate key column to the destination table. Use it when the source has no reliable key.
Load Strategy How the table is loaded. See Choose a load strategy.
Destination Table Click the name or the pencil to rename it.

The destination cell also shows badges:

  • N columns mapped — included columns.
  • N to review — AI name suggestions waiting for you.
  • No PK — no included column is marked as a key.
  • N calculated — calculated columns that need a handling decision. See Override a mapping with an expression.

Column mappings

Click a table row to expand its columns. Only one table expands at a time.

Column What it does
Include checkbox Clear it to leave the column out.
Source Column, Source Type The source column and its data type. A Stale tag means the column no longer exists in the source.
ƒ icon The source or destination column is calculated. Hover for details.
Destination Column Click the name or the pencil to rename it.
Destination Type The destination data type. Read-only here.
Strategy Appears for calculated columns only. See Override a mapping with an expression.
Is PK Marks the column as part of the destination key. Most load strategies need at least one.

Destination names are styled by how they were set:

  • Plain — from your naming conventions.
  • Blue — an AI suggestion you accepted.
  • Bold — you renamed it yourself.

Rename a table or column

  1. Click the destination name or its pencil icon.
  2. Type the new name.
  3. Press Enter or click away to save. Press Esc to cancel.

Review AI suggestions

When the AI rename has run, columns with a suggestion show a line such as Suggest: CustomerId (92%) under the destination name.

  • Accept applies the suggested name.
  • Reject discards the suggestion and keeps the current name.

To run the AI rename, open the Documentation tab.

Documentation tab

Holds the same Source System Documentation and Additional Instructions fields as the create dialog. Your edits save when you click out of a field.

To ask for name suggestions:

  1. Add or update the documentation. Run AI Rename needs documentation. Without it you see Please add documentation first.
  2. Click Run AI Rename.
  3. Wait for the Running AI Rename dialog to finish. It reports progress per table.
  4. Review the suggestions on the Table Mappings tab.

The AI rename counts toward your organization’s AI allowance. If the allowance is used up, it does not start. See AI in Pipelines.

Summary tab

A read-only snapshot of the Pipeline:

  • Pipeline Overview — sources, destination and its database system, status, and created date.
  • Mapping Statistics — Tables Included, Tables Excluded, Total Columns.
  • Columns Renamed — counts by Convention, AI Accepted, and Manual, plus AI Suggestions Generated and Pending Review.
  • Views Being Materialized — each source view that becomes a destination table, and whether it has a key (PK designated or No PK designated).
  • Data Type Conflicts — type changes between source and destination, labelled as a precision difference, type conversion, or lossy conversion.

Editing a finalized Pipeline

You can keep editing after you finalize. Changing names, inclusion, keys, strategies, or calculated-column handling sets the status back to Mapping In Progress and marks the affected load procedures as out of date. The destination schema does not change until you validate and finalize again.