MySQL

SQL

The MySQL connector is a SQL connector that queries a MySQL database and maps rows to claims using a column mapping dictionary.

How it works

Uses fetch() to yield claim dicts; the base class run() handles batching and ingestion
Connects via pymysql with DictCursor for named column access
Values are automatically converted to strings before claim construction
Supports MariaDB-compatible servers — any server that speaks the MySQL wire protocol
Provenance is tagged as mysql:{query} with source type database_import

Column mapping

The mapping dict tells the connector which columns correspond to claim fields. At minimum you need subject, predicate, and object keys. Optionally include confidence to map a numeric column to claim confidence.

Mapping key	Required	Description
`subject`	Yes	Column name for subject entity
`predicate`	Yes	Column name for predicate/relation
`object`	Yes	Column name for object entity
`confidence`	No	Column name for numeric confidence value

1 Install pymysql

The MySQL connector requires the pymysql package for database access.

$ pip install pymysql

2 Prepare your database

You need the hostname, username, password, and database name for your MySQL server. The default port is 3306. Ensure your user has SELECT privileges on the tables you plan to query.

3 Write a query

Write a SQL query that returns columns for subject, predicate, and object. The connector runs the query and maps each row to a claim using your mapping dict.

4 Connect

conn = db.connect("mysql",
    host="localhost",
    user="root",
    password="secret",
    database="mydb",
    query="SELECT src, rel, dst FROM edges",
    mapping={"subject": "src", "predicate": "rel", "object": "dst"},
)
result = conn.run(db)

Parameter	Required	Default	Description
`host`	Yes	—	MySQL server hostname
`user`	Yes	—	Database username
`password`	Yes	—	Database password
`database`	Yes	—	Database name
`query`	Yes	—	SQL query to execute
`mapping`	Yes	—	Column-to-claim field mapping dict
`port`	No	`3306`	MySQL port
`source_type`	No	`database_import`	Provenance source type
`save`	No	`False`	Encrypt and persist credentials

Basic usage

conn = db.connect("mysql",
    host="localhost",
    user="root",
    password="secret",
    database="mydb",
    query="SELECT src, rel, dst FROM edges",
    mapping={"subject": "src", "predicate": "rel", "object": "dst"},
)
result = conn.run(db)

DSN-style connection (alternative)

conn = db.connect("mysql",
    dsn="mysql://user:pass@host/db",
    query="SELECT gene, relation, target FROM interactions",
    mapping={"subject": "gene", "predicate": "relation", "object": "target"},
)
result = conn.run(db)

With save and custom port

conn = db.connect("mysql",
    host="db.example.com",
    user="reader",
    password=os.environ["MYSQL_PASS"],
    database="prod",
    port=3307,
    query="SELECT a, b, c FROM data",
    mapping={"subject": "a", "predicate": "b", "object": "c"},
    save=True,
)
result = conn.run(db)