DB Push

The DB Push node writes a DataFrame directly to a database table using automatic column mapping from the input DataFrame schema. Unlike DB Table Push, it does not use a visual mapper.

Overview

Use DB Push for straightforward database writes where the DataFrame columns should become the table columns, or where an existing table already matches the DataFrame field names.

Purpose

• Create a destination table from the input DataFrame schema when the table does not exist.
• Insert or update rows using the configured primary key fields.
• Push large DataFrames in batches, with optional staging-table merge behavior.

Configuration

Field	Required	Description
Table Name	Yes	Destination table name. Spaces are excluded by the form control.
Table Description	Yes	Description/comment applied when a new table is created.
Primary Keys	Yes	One or more input columns used as the primary key and upsert key.
Schema	Yes	Target database schema connection.
Batch Size	Yes	Number of rows submitted per database batch. Default configuration value is 1000.
Use Staging Table	No	Enables staging-table merge for larger inputs.
Variation Field	Required when staging is enabled	Field used by staging merge logic to resolve duplicates.
Threshold	Required when staging is enabled	Row-count threshold that activates staging mode. Default configuration value is 2000.

Inputs

Input	Data Type	Input Type	Description
DataFrame	DataFrame	Single	The data to write to the destination database table.

Outputs

Output	Data Type	Collection	Description
Output	DataObject	False	Result JSON containing {"result":"success"} after a successful write.

Properties

Property	FlowDataType	Description
TARGET_DB_NAME	String	Target database name
TARGET_SCHEMA_NAME	String	Target schema name
TARGET_TABLE_NAME	String	Target table name

Processing Logic

• Loads the selected database schema and opens a database connection.
• Validates that Variation Field is set when Use Staging Table is enabled.
• Checks whether the destination table exists.
• If the table does not exist, creates it from the input DataFrame schema and configured primary keys. Column types are inferred from the input schema and adjusted for the target database dialect.
• Reads DataFrame data from workflow storage and writes it in batches.
• For normal mode, prepares upsert statements using the primary keys.
• For staging mode, creates a staging table when the deduplicated input row count reaches the threshold, writes to staging, merges staging into the destination table, then drops the staging table.
• Aborting the node stops the database update process.

Examples

Create or update an orders table:

• Table Name: orders
• Primary Keys: order_id
• Schema: ReportingDB\public
• Batch Size: 1000

Notes / Limitations

• DataFrame field names drive destination column names. For an existing table, the table columns must match the input schema expected by the node.
• The node may create a new table if the configured table name does not exist; ensure the database user has create-table privileges.
• Staging mode creates a temporary table named after the destination table with a _nazdaq_staging suffix (upper-cased for Oracle), then drops it after merge.
• A batch size of 0 fails validation at runtime.
• For explicit field mapping, static values, scripts, or custom destination schemas, use DB Table Push.