Skip to main content

Load node

Target

The Load node writes rows to a target system: database table (insert / upsert / update / delete), a file at a file-transfer target (CSV, JSON, Parquet), an HTTP API (POST / PUT), or a SaaS target (Salesforce bulk API, Google Sheets). Load is the terminal step for most flows.

When to use​

  • The workflow's job is to move data into a system of record: Load is where the write happens.
  • You need per-connection semantics (bulk insert vs single insert, upsert on a key, transactional writes).
  • The write pattern matches a first-class connector; when it doesn't, Message + DB Script gives a two-step "format then write" alternative.

Ports​

PortDirectionKindNotes
inindataRows to write.
outoutdataSummary row per chunk (rows written, rows failed, target reference).
erroroutcontrolFires on write failure.
notifyoutcontrolNotification policy port.

Configuration​

  • Connection. Any write-capable connection (SQL family, Salesforce, S3, Drive, SharePoint, Dropbox, SFTP, Generic API).
  • Action. A bound Saved Action or in-node action config. Same shape as the Source node.
  • Action operation. insert, upsert, update, delete for SQL / Salesforce; Upload, Put file, Append for file transfer; POST, PUT for Generic API.
  • Target object / path / URL. The specific target: table name, object, remote path, URL suffix.
  • Key columns (upsert / update / delete). The columns that identify a target row.
  • Batch size. Overrides the runtime's memory-adaptive chunk size for writes.

Configure Action walkthrough​

  1. System family — pick the connector.
  2. Connection — pick the connection.
  3. Action — Saved Action or ad-hoc. Preview runs against the connection.
  4. Field mapping — maps input row columns to target columns. Auto-suggested when the target schema is known.
  5. Advanced — batch size, retry policy, transactional mode, pre/post hook actions.

Runtime behaviour​

  • The runtime writes per-chunk, so each upstream chunk fires one write. The chunker respects batchSize when set.
  • SQL family writes go through the same connector adapters Source and DbScript use, with per-dialect bulk paths.
  • Salesforce writes route through the bulk API when the chunk crosses the bulk threshold.
  • File writes go through the connector's write shape (putFile, uploadPart).

Failure modes​

  • Constraint violation. Row-level failure lands on the error port with the offending row identity.
  • Auth expired. The connector's refresh path fires; if it still fails the connection needs re-auth on the Connections page.
  • Target full / quota exceeded. Depends on the target — S3 returns a rate-limit code, SQL Server may throw a specific error class. Surface on error.

Source

Read rows from a source system

Transform

Map and derive fields before write

DB Script

Prep the target with DDL / cleanup