AWS Glue Reference

Reference material for AWS Glue Publishing -- data type mappings, CLI arguments, and querying.

Data Type Mapping

Source types map automatically to Glue-compatible types. LakeXpress exports through Parquet, so the mapping is driven by the normalized source type name.

Pick your source database:

PostgreSQL

PostgreSQL to Glue

PostgreSQL Type	Glue Type
SMALLINT	smallint
INTEGER, INT4	int
BIGINT, INT8	bigint
NUMERIC(p,s)	decimal(p,s)
REAL, FLOAT4	float
DOUBLE PRECISION, FLOAT8	double
VARCHAR(n), TEXT, CHAR(n), TIME, UUID, JSON, JSONB, array []	string
BOOLEAN	boolean
DATE	date
TIMESTAMP, TIMESTAMPTZ	timestamp
BYTEA	binary

SQL Server

SQL Server to Glue

SQL Server Type	Glue Type
TINYINT	tinyint
SMALLINT	smallint
INT	int
BIGINT	bigint
DECIMAL(p,s)	decimal(p,s)
MONEY	decimal(19,4)
SMALLMONEY	decimal(10,4)
FLOAT, REAL	float
BIT	boolean
VARCHAR(n), NVARCHAR(n), CHAR(n), NCHAR(n), TEXT, NTEXT, UNIQUEIDENTIFIER	string
DATE	date
DATETIME, DATETIME2, SMALLDATETIME, DATETIMEOFFSET	timestamp
VARBINARY	binary

Oracle

Oracle to Glue

Oracle Type	Glue Type
NUMBER	decimal(38,18)
NUMBER(p,s)	decimal(p,s)
FLOAT	float
VARCHAR2(n), NVARCHAR2(n), CHAR(n), CLOB, NCLOB	string
DATE	date
TIMESTAMP	timestamp
RAW, BLOB	binary

Teradata

Teradata to Glue

Teradata Type	Glue Type
BYTEINT, VARCHAR(n), CHAR(n), CLOB, TIME, VARBYTE	string
SMALLINT	smallint
INTEGER	int
BIGINT	bigint
DECIMAL(p,s), NUMERIC(p,s)	decimal(p,s)
NUMBER	decimal(38,18)
FLOAT, REAL	float
DATE	date
TIMESTAMP	timestamp
BYTE, BLOB	binary

MySQL

MySQL to Glue

MySQL Type	Glue Type
TINYINT	tinyint
SMALLINT	smallint
MEDIUMINT, YEAR, VARCHAR(n), CHAR(n), TINYTEXT, TEXT, MEDIUMTEXT, LONGTEXT, ENUM, SET, TIME, TINYBLOB, MEDIUMBLOB, LONGBLOB, JSON	string
INT	int
BIGINT	bigint
DECIMAL(p,s)	decimal(p,s)
FLOAT	float
DOUBLE	double
BIT	boolean
DATE	date
DATETIME, TIMESTAMP	timestamp
BINARY, VARBINARY, BLOB	binary

MariaDB

MariaDB to Glue

MariaDB Type	Glue Type
TINYINT	tinyint
SMALLINT	smallint
MEDIUMINT, YEAR, VARCHAR(n), CHAR(n), TINYTEXT, TEXT, MEDIUMTEXT, LONGTEXT, ENUM, SET, TIME, TINYBLOB, MEDIUMBLOB, LONGBLOB, JSON	string
INT	int
BIGINT	bigint
DECIMAL(p,s)	decimal(p,s)
FLOAT	float
DOUBLE	double
BIT	boolean
DATE	date
DATETIME, TIMESTAMP	timestamp
BINARY, VARBINARY, BLOB	binary

SAP HANA

SAP HANA to Glue

SAP HANA Type	Glue Type
TINYINT	tinyint
SMALLINT	smallint
INTEGER	int
BIGINT	bigint
DECIMAL(p,s)	decimal(p,s)
SMALLDECIMAL, VARCHAR(n), NVARCHAR(n), CHAR(n), NCHAR(n), CLOB, NCLOB, TEXT, TIME, SECONDDATE	string
REAL	float
DOUBLE	double
BOOLEAN	boolean
DATE	date
TIMESTAMP	timestamp
VARBINARY, BLOB	binary

CLI Reference

Glue Publishing Arguments

Option	Type	Description
--publish_target ID	String	Credential ID for Glue publishing (required)
--publish_schema_pattern PATTERN	String	Glue database naming pattern (default: EXT_{schema})
--publish_database_name NAME	String	Literal Glue database name; overrides --publish_schema_pattern when set
--publish_table_pattern PATTERN	String	Table naming pattern (default: {table})
--n_jobs N	Integer	Parallel workers for table creation (default: 1)

There is no --glue_skip_existing flag -- AWS Glue publishing always (re)creates tables.

Querying Glue Tables

Amazon Athena:

SELECT * FROM lx_tpch_1.customer LIMIT 10;

AWS Redshift Spectrum:

SELECT * FROM spectrum_schema.customer LIMIT 10;

Amazon EMR (Spark):

df = spark.table("lx_tpch_1.customer")
df.show(10)

See Also

• AWS Glue Publishing - Setup and usage guide
• CLI Reference - All command-line options