cnxload

name	cnxload
description	Load files (Parquet/CSV/JSON) into a JDBC table

Connection Load Skill

Loads data from Parquet, CSV, or JSON files directly into a JDBC database table. This provides a simple way to push data into any JDBC-compatible database without going through the full Starlake load pipeline.

Usage

starlake cnxload [options]

Options

• --source_file <value>: Full path to the source file (required). Supports Parquet, CSV, JSON
• --output_table <value>: Target JDBC table in schema.table format (required)
• --options k1=v1,k2=v2: JDBC connection options:
  • driver: JDBC driver class (e.g. org.postgresql.Driver)
  • user: Database user
  • password: Database password
  • url: JDBC URL (e.g. jdbc:postgresql://localhost:5432/mydb)
  • partitions: Number of partitions for parallel write
  • batchSize: Batch size for inserts
• --write_strategy <value>: Write strategy: APPEND, OVERWRITE
• --reportFormat <value>: Report output format: console, json, or html

Examples

Load Parquet to PostgreSQL

starlake cnxload \
  --source_file /data/orders.parquet \
  --output_table public.orders \
  --options driver=org.postgresql.Driver,url=jdbc:postgresql://localhost:5432/mydb,user=admin,password=secret \
  --write_strategy APPEND

Load CSV to Database

starlake cnxload \
  --source_file /data/customers.csv \
  --output_table public.customers \
  --options driver=org.postgresql.Driver,url=jdbc:postgresql://localhost:5432/mydb,user=admin,password=secret \
  --write_strategy OVERWRITE

Load with Batch Options

starlake cnxload \
  --source_file /data/orders.parquet \
  --output_table sales.orders \
  --options driver=com.mysql.cj.jdbc.Driver,url=jdbc:mysql://localhost:3306/mydb,user=root,password=secret,batchSize=5000 \
  --write_strategy APPEND

Related Skills

• load - Full Starlake load pipeline
• esload - Load data into Elasticsearch
• kafkaload - Load/offload data to/from Kafka