Data Output

Data Output exports incoming data when the pipeline runs. It can start a local download, save a file to supported object storage, or write rows to a Snowflake table. The node accepts exactly one upstream connection. It is terminal in the workflow UI and cannot connect downstream.

Basic Usage

1. Add Data Output to the canvas and connect the data you want to export.
2. Choose Download Locally or select a configured destination.
3. For a local or object-storage export, choose CSV or Excel. For object storage, also enter a base filename.
4. For Snowflake, choose a write mode. An upsert also requires a primary-key column.
5. Select Apply, then run the pipeline. Apply saves the settings; running the pipeline performs the export.

Destinations

Download Locally

Download Locally is the default. After a successful run, Rhombus prepares the output and starts a browser download in one of these formats:

• CSV (.csv)
• Excel (.xlsx)

The filename is the node's execution name followed by the selected extension. Custom Filename does not change the name of a local browser download.

Object Storage

You can create and select destinations for:

• Amazon S3
• Azure Blob Storage
• Google Cloud Storage

Object-storage destinations use the selected CSV or Excel format. Add and save credentials through Add New Destination. The destination must belong to the current project.

Configure an Amazon S3 destination

Before adding Amazon S3 in the node, create a dedicated, least-privilege AWS identity that can write to the output bucket. The destination setup requires an AWS access key ID, secret access key, the bucket's actual Region, and the bucket name.

The identity must be able to inspect the bucket and write output objects. For large exports, allow multipart-upload cleanup as well. See Configure Amazon S3 for Data Output for the AWS Console steps, example IAM policy, SSE-KMS permissions, and troubleshooting guidance.

An Amazon S3 source connection is read-only and does not configure this destination. Keep source-read and output-write permissions separate.

Snowflake

A Snowflake destination writes the incoming DataFrame directly to its configured table, so the file format and filename controls are hidden. Choose a write mode:

• Append: Add incoming rows to the table. This is the default.
• Overwrite: Replace the table contents with the current output.
• Upsert (Merge): Match rows by the selected primary-key column, update matches, and insert non-matches. A primary key is required.

Rhombus can create the destination table when needed. If the upstream schema is available, the panel can also generate optional setup SQL.

File Naming

For object-storage exports, Custom Filename is a base name without an extension. Rhombus adds a millisecond timestamp and the selected extension to avoid duplicate names. For example, clean_sales becomes clean_sales_<timestamp>.csv.

If the field is empty, Rhombus uses RhombusAI_output. A custom base name:

• may contain letters, numbers, hyphens, and underscores;
• must start and end with a letter or number;
• cannot contain spaces, periods, special characters, or consecutive hyphens; and
• must contain no more than 255 characters.

Standard and Big Data Execution

In Standard execution, a local download contains the node's output. In Big Data/AWS Glue execution, it contains only the node's 100-row preview snapshot, not the full distributed result. Use a remote destination to write the complete output from a Big Data run.

Troubleshooting

• Apply is unavailable: Select Download Locally or a valid project destination.
• Upsert cannot be applied: Select a primary-key column.
• Snowflake setup SQL is unavailable: Connect and run the upstream nodes so Rhombus can infer their output schema.
• The expected file is not present yet: The pipeline run performs the export; applying the node does not.
• A local download fails: After the run, use the node's manual download action and check the workflow error log.

See Data Input to configure the dataset and sample that enter the workflow.