Azure Synapse Analytics Nodes

Azure Synapse Analytics is Microsoft's analytics service: a serverless SQL pool that queries the data lake, dedicated SQL pools for warehousing, and pipelines for orchestration. These nodes run T-SQL against a pool, land pipeline output in a dedicated pool, move data between the pools and Azure Storage, read the catalog, and start and follow Synapse pipeline runs.

Set the connection up first: see the Azure Synapse Analytics Integration Guide.

Configuration Quick Reference

Field	What you choose	Details
Parameters	Connection, Function, Function Parameters, Timeout Override	Select the connection profile and function, bind the function's parameters to upstream values or constants, and optionally override the timeout.
Settings	Description, Timeout (seconds), Retry on Timeout, Retry on Fail, On Error	Node description, maximum execution time, retry behaviour, and error handling. All execution settings default to the pipeline's.

The pool is a property of the connection, not the node

A node binds to a connection, and the connection decides which SQL pool statements run on. Write and Copy Into need a connection to a dedicated pool; on a serverless connection they are refused before anything is sent.

Zoom

Azure Synapse Query Node

Azure Synapse Query Node

Run a SELECT — including OPENROWSET over lake files on the serverless pool — and pass the rows downstream.

Function Name	Purpose	Common Use Cases
Query	Run a parameterised SELECT	Warehouse aggregates, lake queries, enrichment lookups, data-quality checks

{
  "rows": [{ "machine_id": "press-01", "readings": 8, "avg_temp": 40.76 }],
  "rowCount": 1
}

Read a value downstream as $node["Query"].result.rows[0].machine_id.

Zoom

Azure Synapse Execute Node

Azure Synapse Execute Node

Run a statement that changes data or schema and report how many rows it touched.

Function Name	Purpose	Common Use Cases
Execute	Run DML or DDL	Housekeeping deletes, UPDATE STATISTICS, CTAS rebuilds, DDL before a load

{ "rowsAffected": 3 }

rowsAffected is -1 when the statement has no row count of its own.

Zoom

Azure Synapse Write Node

Azure Synapse Write Node

Insert pipeline records into a dedicated-pool table, deriving the columns from the batch.

Function Name	Purpose	Common Use Cases
Write	Insert records with schema detection	Landing OT telemetry, persisting transform output, appending to a fact table

{ "rowsInserted": 3 }

Zoom

Azure Synapse Copy Into Node

Azure Synapse Copy Into Node

Bulk-load CSV, Parquet or ORC files from Azure Storage into a dedicated-pool table with COPY INTO.

Function Name	Purpose	Common Use Cases
Copy Into	Load files the pool reads in parallel	Ingesting a day of landed files, backfilling from the lake

{ "table": "sensor_archive", "sourceUrl": "https://lake.dfs.core.windows.net/raw/readings/2026-09-28/*.csv", "rowsLoaded": 48 }

The pool reads the files itself, with the connection's storage credential; the rows never pass through the pipeline.

Zoom

Azure Synapse Export To Storage Node

Azure Synapse Export To Storage Node

Write a query's result to the lake as a new external table (CREATE EXTERNAL TABLE AS SELECT).

Function Name	Purpose	Common Use Cases
Export To Storage	Export a result set as Parquet or CSV files	Handing an extract to another system, archiving a slice of history

{ "externalTable": "readings_2026_09", "location": "exports/2026-09/", "rowsExported": 48 }

Zoom

Azure Synapse List Tables Node

Azure Synapse List Tables and Describe Table Nodes

Read the catalog: the tables, views and external tables in a schema, and one table's columns.

Function Name	Purpose	Common Use Cases
List Tables	List a schema's tables, optionally filtered by a LIKE pattern	Discovering what is queryable, looping over a set of tables
Describe Table	Read a table's column definitions	Driving a mapping step, checking a target before a write

{ "tables": [{ "schema": "dbo", "name": "sensor_history", "type": "BASE TABLE" }], "count": 1 }

{
  "schema": "dbo",
  "table": "sensor_history",
  "columns": [{ "name": "machine_id", "type": "nvarchar(4000)", "nullable": true }],
  "columnCount": 1
}

A table that does not exist is a permanent failure for Describe Table — asking again returns the same nothing.

Zoom

Azure Synapse Run Pipeline Node

Azure Synapse Pipeline Nodes

Start a Synapse pipeline, follow it, cancel it, and list what can be run. These need a connection that signs in with Microsoft Entra ID.

Function Name	Purpose	Common Use Cases
Run Pipeline	Start a run and return its run ID at once	Refreshing a mart after files land, re-running a backfill for one day
Get Pipeline Run	Read a run's status, timings and failure message	Waiting for a run to finish, branching on its outcome
Cancel Pipeline Run	Ask Synapse to cancel a queued or running run	Stopping a run started with the wrong parameters
List Pipelines	List published pipelines with their parameters	Finding a pipeline's parameter names before wiring Run Pipeline

{ "pipelineName": "refresh_sales_mart", "runId": "a37adbaa-db2f-44fc-be75-41ad12fc86c2" }

Run Pipeline does not wait. To act on the outcome, loop a Get Pipeline Run node — its runId bound to $node["Run Pipeline"].result.runId — until $node["Get Pipeline Run"].result.terminal is true, then branch on result.succeeded. A failed run is a successful read: the node succeeds, and result.message carries Synapse's reason.

Retry on Fail can start a second run

Run Pipeline sends its request once. If the call times out or the connection drops after Synapse received it, the run may have started while the node reports a failure. With Retry on Fail turned on, the retry starts a second run. Leave it off on Run Pipeline, and check the workspace's Monitor hub before starting the run again by hand.

Output

Every Synapse node delivers its data under result, and execution facts (success, functionId, durationMs, timestamp) under _metadata.

Node	Expression	Description
Query	$node["Name"].result.rows	The result rows, one object per row keyed by column name
	$node["Name"].result.rowCount	How many rows were delivered
Execute	$node["Name"].result.rowsAffected	Rows the statement affected; -1 for a statement with no count
Write	$node["Name"].result.rowsInserted	Rows inserted across every batch of this execution
Copy Into	$node["Name"].result.rowsLoaded	Rows COPY INTO loaded
	$node["Name"].result.table	The table it loaded into
	$node["Name"].result.sourceUrl	The storage location it read
Export To Storage	$node["Name"].result.rowsExported	Rows written to the lake
	$node["Name"].result.externalTable	The external table created over the files
	$node["Name"].result.location	The folder the files were written to
List Tables	$node["Name"].result.tables	One object per table, each with schema, name and type (BASE TABLE, VIEW or EXTERNAL TABLE)
	$node["Name"].result.count	How many tables matched
Describe Table	$node["Name"].result.columns	One object per column, each with name, the declared type and nullable
	$node["Name"].result.columnCount	How many columns the table has
	$node["Name"].result.schema	The schema it read from
	$node["Name"].result.table	The table it described
Run Pipeline	$node["Name"].result.runId	The run ID Synapse assigned
	$node["Name"].result.pipelineName	The pipeline that was started
Get Pipeline Run	$node["Name"].result.status	Queued, InProgress, Succeeded, Failed, Canceling or Cancelled
	$node["Name"].result.terminal	Whether the run has stopped for good
	$node["Name"].result.succeeded	Whether it finished successfully
	$node["Name"].result.message	Why it failed; empty otherwise
	$node["Name"].result.runStart, result.runEnd	RFC 3339 times; runEnd is empty while the run is going
	$node["Name"].result.durationMs	How long it ran; 0 while it is going
	$node["Name"].result.lastUpdated	When Synapse last updated the run
	$node["Name"].result.parameters	The parameter values the run was started with
	$node["Name"].result.runGroupId, result.isLatest	The recovery group the run belongs to, and whether it is the latest in it
	$node["Name"].result.invokedBy	What started it: Manual for an API call, or a trigger's name
	$node["Name"].result.runId, result.pipelineName	The run and its pipeline
Cancel Pipeline Run	$node["Name"].result.runId	The run Synapse accepted a cancel request for
	$node["Name"].result.recursive	Whether the runs it started were asked to cancel too
List Pipelines	$node["Name"].result.pipelines	One object per pipeline: name, description, folder, parameters and activityCount
	$node["Name"].result.count	How many were listed
	$node["Name"].result.truncated	Whether more exist than Max Items allowed

A Query node's picker offers the connection's Execute and Write functions too, so a Query node bound to one of those delivers that function's key instead: result.rowsAffected or result.rowsInserted.

The call's own facts, under _metadata

Every node carries _metadata.method, _metadata.connectionId, _metadata.protocol (always azuresynapse) and _metadata.pool — serverless or dedicated, the pool the connection points at.

A Query adds _metadata.driver (always azuresynapse), _metadata.rowCount, _metadata.truncated (true when a row cap cut the result short) and _metadata.columns, each result column's name, type and nullability. An Execute adds _metadata.driver alone.

A Write adds _metadata.driver, _metadata.table, _metadata.schema, _metadata.batchSize, _metadata.totalRows and _metadata.matchedColumns (the table columns the data was mapped onto). It adds _metadata.skippedFields when the data carried fields the table has no column for, and _metadata.schemaEvolution when the write added columns. A write that created the table first reports _metadata.tableCreated and the _metadata.columns it created. A write that failed part-way reports _metadata.rowsInserted, the rows that landed before the failure.

Copy Into adds _metadata.table and _metadata.sourceUrl; Export To Storage adds _metadata.externalTable and _metadata.location. The two catalog nodes add _metadata.schema, and Describe Table adds _metadata.table. Run Pipeline adds _metadata.pipelineName; Get Pipeline Run and Cancel Pipeline Run add _metadata.runId.

Check truncated before you aggregate

A query stopped at the row cap returns a shorter rows array that looks exactly like a complete result. The cap is the connection's Max Result Rows (1000 by default). Branch on _metadata.truncated, or narrow the query, rather than assuming the read was complete.

Store and Forward

Write and Copy Into are store-and-forward eligible. With durability enabled on the binding, a batch that cannot reach the pool — a paused pool included — is buffered and drained when it recovers, rather than lost.

Both are non-idempotent: a replayed batch inserts the rows again, and a replayed COPY INTO loads the same files again. See Store & Forward for the full contract.

Error Handling

Kind of failure	Classification	Examples
The statement, the schema or the setup	Permanent — straight to the DLQ	Syntax error, unknown object, value too wide for its column, a number with a fraction for a whole-number column, login failed, permission denied, firewall refusal, missing files for Copy Into, an unknown pipeline, a pipeline RBAC refusal
The service, the network, or concurrency	Transient — Write and Copy Into are buffered and retried; the other nodes fail and follow their Retry settings	Connection lost, deadlock, a pool that is paused or warming up, Azure SQL's "retry later" errors, HTTP 429 and 5xx from the development endpoint

Node errors carry the pool's own message and error number, for example Invalid object name 'dbo.readings'. (error 208), and Entra ID refusals are cut down to their reason.

Related Nodes

• Microsoft SQL Server nodes — the same T-SQL dialect on an operational database
• Amazon Redshift nodes, Snowflake nodes and BigQuery nodes — the other cloud warehouses