Projections

A projection is a managed data workflow that publishes a curated or consumption-ready dataset from selected source entities and attributes into a target entity. Use the Projections screen to define what to publish, whether it refreshes in real time (Live) or on a schedule (Batch), and to monitor, pause, or resume each projection’s underlying job.

Projections are the last hop of the pipeline: they take entities Nexa has already curated and materialize a business-facing dataset (all projections here write to the consumption layer). Creating and editing projections is restricted to the dev environment — the Add, Edit, and Delete controls only appear there.

The Projections list

Each row represents one projection:

Column	Meaning
Projection Name	The projection’s name
Alias	Its generated alias
Source Entities	Source entity pills; the first two show, then a +N more expander
Target Layer	curated (red badge) or consumption (blue badge)
Type	live (green badge) or batch (blue badge)
Status	The projection state (see below)
Actions	Play/pause, view, edit, and delete controls

Status values

Status	Badge
running	Green
completed	Green
suspended	Yellow
paused	Yellow
pending	Yellow
failed	Red

Filtering

Open Filter to narrow the list:

• Projection Type — All types, Live, or Batch.
• Projection Mode — All Modes, Legacy Projections, or Non-legacy projections.
• Sort By — Created Date, Last Modified, or Projection Name, ascending or descending.

Search by projection name with the search box.

Row actions

• Play/Pause — pauses or activates the projection’s job (shown only when the projection has a job).
• View — opens the underlying job panel with its Jobs Details and Runs tabs, so you can inspect run history.
• Edit — reopens the create form in edit mode (dev only).
• Delete — prompts Are you sure you want to delete this projection? before removing it (dev only).

Create a projection

Select Add to open the Create New Projection panel.

1. Choose the projection style. Set Legacy Projection to True or False. Choose True to generate mappings from an uploaded C360 (Customer 360) model file; choose False to define a standard projection by hand.

2. Name and describe it (standard projections). Enter a Projection Name (up to 100 characters) and a Projection Description (up to 250 characters). For a legacy projection, upload the model file instead under Upload Legacy Model (JSON only) — one projection is created per table in the model.

3. Select the projection type. Pick one:

   • Live — real-time data processing. Selecting Live clears any schedule.
   • Batch — scheduled data processing. Selecting Batch reveals the schedule builder.

4. Set the schedule (Batch only). Use the schedule builder to define a cron expression (see below). You must save a schedule before you can continue.

5. Select the source. Under Select Source, choose entities or attributes, grouped by entity. Selecting a group header toggles all of its attributes. ETL (extract, transform, load) logic for these sources is authored on the Canvas.

6. Name the target. Enter a target entity name under Enter Target Entity Name.

7. Set options and save. Optionally enable PII Suppressed to exclude PII (personally identifiable information) from the projection, then select Create Projection. Editing an existing projection shows Update Projection instead.

Batch schedules

For a batch projection, the schedule builder produces a cron expression from plain-language options and shows a live preview (for example, Cron: 0 30 * * * with the readable line “Every hour at :30 past the hour”). The schedule saves automatically as you edit it — there is no separate save button.

Frequency	Options you set
Every minute	Interval in minutes
Hourly	Interval in hours, minute of the hour, starting hour
Daily	Every day or every weekday, and start time
Weekly	Days of the week and start time
Monthly	Day of month or weekday pattern, month interval, and start time

Batch Projection Settings

The separate Batch Projection Settings screen manages older batch configurations saved locally in your browser (schedule, output format, compression, and per-entity attributes). It is independent of the projections above; edit those saved configurations from the Consumption files area. New projections should be created from the Projections screen.

Live vs. batch

Databricks

A Live projection refreshes continuously as source data changes; a Batch projection runs on its cron schedule. Both are backed by jobs you can monitor from the View panel and from the Jobs screen.

Snowflake

Projections work the same way; Live and Batch map to the equivalent Snowflake refresh and scheduled-task constructs.