Overview

The Output node is the destination node in a Weave Workflow. It writes the data produced by upstream nodes to either a Datasource (database table) or a Connector (file in CSV or Excel format). Output nodes are terminal — they accept upstream input but emit nothing to further downstream nodes. A Workflow typically ends at one or more Output nodes; multiple Output nodes can fan out the same upstream data to different destinations.

The node exposes four write modes via radio buttons — Override, Append, Update, Upsert — that control how the incoming data interacts with existing data at the destination. Update and Upsert require a primary key for row identification; Update additionally requires column selection. File destinations support max-records-per-file splitting and date-expression file naming.

Note: “Workflow” is the in-UI term for what some Weave documentation calls a pipeline. This reference uses “Workflow.”

When to use it

• Writing a Workflow’s processed output to a database table for downstream Reports, applications, or further pipelines to consume.
• Exporting transformed data to a file (CSV / Excel) for distribution, backup, or external system consumption.
• Synchronizing data between systems where the Workflow extracts from one source and loads into another.
• Maintaining incrementally updated tables — Update for known-row modification, Upsert for mixed update-and-insert.
• Refreshing aggregated or reference tables on schedule via Override.

Role in the Workflow

Upstream inputs	One input. Accepts data from a Data Import, transformation, or other upstream node.
Downstream outputs	None. Output is a terminal node — its work is writing to a destination, not emitting data forward.
Destination	Either a Datasource (database table) or a Connector (file).
Run trigger	Manual via the node’s Run icon for testing, automatic when the Workflow runs as a whole.

Write modes

The four write modes control how the incoming data interacts with existing data at the destination. Pick exactly one per Output node.

Override	Truncates the existing destination table and writes the current Workflow’s data. Destructive — existing data at the destination is replaced.	Each Workflow run produces a complete replacement of the destination (e.g., daily full-table refreshes, reference data updates).
Append	Adds new rows below existing rows in the destination. Existing rows are untouched.	Each Workflow run produces new records to accumulate (e.g., log streams, transaction inserts, time-series data).
Update	Modifies selected columns of existing rows identified by a primary key. Rows in the upstream that don’t match an existing primary key are not written.	Each Workflow run produces updated values for known records (e.g., recalculated metrics on existing customers, status updates on existing orders).
Upsert	Combines Update and Insert: rows that match the primary key are updated; rows that don’t match are inserted as new records.	Each Workflow run produces a mix of new and existing records (e.g., daily syncs where some entities are new and some have changes).

Destination types

Each Output node writes to one destination — either a Datasource (database) or a Connector (file). The configuration UI differs by destination type.

Data Source destination

Data Source	The configured Datasource the table is written to.
Dataset Name	The destination table’s name (User given) in the Datasource.
Primary Key (Update, Upsert)	Column or combination of columns uniquely identifying rows. Required for Update and Upsert. Multiple columns can be selected to form a composite key.
Update Columns (Update, Upsert)	Columns to modify on matched rows. Selected via a column-selection popup.

Connector destination

Connector	The configured Connector the files are written to.
File Type	Output file format.
Folder Name	Destination directory on the Connector.
File Name	Destination file name. Supports date expressions in square brackets.
Max Records Per File	Splits output into multiple files when the upstream row count exceeds this limit. Acts as a ceiling — outputs smaller than the limit produce a single file.
Apply (button)	Commits the Connector configuration.

File name date expressions

File names on Connector destinations can include date expressions enclosed in square brackets. The expressions are substituted with the corresponding date / time values at run time.

[YYYY]	Four-digit year (e.g., 2026)
[YY]	Two-digit year (e.g., 26)
[DD]	Day of the month
[H]	Hour
[M]	Minute
[S]	Second

Primary key and column selection (Update and Upsert)

Update and Upsert modes require identifying which rows to modify. This involves two configuration steps.

1. Select primary key	Click the Select button next to the primary key field. A popup appears with the upstream columns. Choose the column or columns whose values uniquely identify rows.
2. Select update columns	Click the Update button. A popup appears. Select the columns that should be modified on matched rows.

Primary key columns must contain unique values. If duplicates exist, row matching is ambiguous and writes produce incorrect results.

Design In View

Output nodes expose a Design In View checkbox. When enabled, the node’s output is available in the design canvas during authoring.

Save / Discard

Configuration follows the Workflow-wide Save / Discard discipline. Changes don’t persist until Save is clicked. Discard reverts to the last saved state. Closing or navigating away without choosing either triggers a warning popup.

Key behaviors

Output is terminal. The Output node accepts upstream input but does not emit data forward. Workflows end at one or more Output nodes; no further nodes can be chained downstream.

Mode selection is exclusive. Override, Append, Update, and Upsert are exclusive radio button choices — exactly one applies per Output node. To write the same upstream data to a destination using different modes (e.g., Append to a log table AND Upsert to a master table), use two Output nodes from the same upstream branch point.

Override is destructive. Override truncates the destination table before writing. Existing data is gone — there is no rollback from within the Output node. For destinations where data must be preserved, use Append, Update, or Upsert.

Fan-out via multiple Output nodes. A Workflow can have multiple Output nodes that all consume from the same upstream point — useful for writing the same processed data to multiple destinations (e.g., a database table AND a CSV file, or a production table AND a backup table).