Databricks Reference

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

Data Type Mapping

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

Pick your source database:

PostgreSQL

PostgreSQL to Databricks

PostgreSQL Type	Databricks Type
SMALLINT	SMALLINT
INTEGER, INT4	INT
BIGINT, INT8	BIGINT
NUMERIC(p,s)	DECIMAL(p,s)
REAL, FLOAT4	FLOAT
DOUBLE PRECISION, FLOAT8	DOUBLE
VARCHAR(n)	VARCHAR(n)
TEXT, TIME, UUID, JSON, JSONB, array []	STRING
CHAR(n)	CHAR(n)
BOOLEAN	BOOLEAN
DATE	DATE
TIMESTAMP	TIMESTAMP_NTZ
TIMESTAMPTZ	TIMESTAMP
BYTEA	BINARY

SQL Server

SQL Server to Databricks

SQL Server Type	Databricks Type
TINYINT	TINYINT
SMALLINT	SMALLINT
INT	INT
BIGINT	BIGINT
DECIMAL(p,s)	DECIMAL(p,s)
MONEY	DECIMAL(19,4)
SMALLMONEY	DECIMAL(10,4)
FLOAT	DOUBLE
REAL	FLOAT
BIT	BOOLEAN
VARCHAR(n), NVARCHAR(n)	VARCHAR(n)
CHAR(n), NCHAR(n)	CHAR(n)
TEXT, NTEXT, UNIQUEIDENTIFIER	STRING
DATE	DATE
DATETIME, DATETIME2, SMALLDATETIME	TIMESTAMP_NTZ
DATETIMEOFFSET	TIMESTAMP
VARBINARY	BINARY

Oracle

Oracle to Databricks

Oracle Type	Databricks Type
NUMBER	DECIMAL(38,18)
NUMBER(p,s)	DECIMAL(p,s)
FLOAT	DOUBLE
VARCHAR2(n), NVARCHAR2(n)	VARCHAR(n)
CHAR(n)	CHAR(n)
CLOB, NCLOB	STRING
DATE	DATE
TIMESTAMP	TIMESTAMP_NTZ
RAW, BLOB	BINARY

Teradata

Teradata to Databricks

Teradata Type	Databricks Type
BYTEINT, 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	DOUBLE
VARCHAR(n)	VARCHAR(n)
CHAR(n)	CHAR(n)
DATE	DATE
TIMESTAMP	TIMESTAMP_NTZ
BYTE, BLOB	BINARY

MySQL

MySQL to Databricks

MySQL Type	Databricks Type
TINYINT	TINYINT
SMALLINT, YEAR	SMALLINT
MEDIUMINT, INT	INT
BIGINT	BIGINT
DECIMAL(p,s)	DECIMAL(p,s)
FLOAT, DOUBLE	DOUBLE
BIT	BOOLEAN
VARCHAR(n)	VARCHAR(n)
CHAR(n)	CHAR(n)
TINYTEXT, TEXT, MEDIUMTEXT, LONGTEXT, ENUM, SET, TIME, JSON	STRING
DATE	DATE
DATETIME, TIMESTAMP	TIMESTAMP_NTZ
BINARY, VARBINARY, BLOB, TINYBLOB, MEDIUMBLOB, LONGBLOB	BINARY

MariaDB

MariaDB to Databricks

MariaDB Type	Databricks Type
TINYINT	TINYINT
SMALLINT, YEAR	SMALLINT
MEDIUMINT, INT	INT
BIGINT	BIGINT
DECIMAL(p,s)	DECIMAL(p,s)
FLOAT, DOUBLE	DOUBLE
BIT	BOOLEAN
VARCHAR(n)	VARCHAR(n)
CHAR(n)	CHAR(n)
TINYTEXT, TEXT, MEDIUMTEXT, LONGTEXT, ENUM, SET, TIME, JSON	STRING
DATE	DATE
DATETIME, TIMESTAMP	TIMESTAMP_NTZ
BINARY, VARBINARY, BLOB, TINYBLOB, MEDIUMBLOB, LONGBLOB	BINARY

SAP HANA

SAP HANA to Databricks

SAP HANA Type	Databricks Type
TINYINT	TINYINT
SMALLINT	SMALLINT
INTEGER	INT
BIGINT	BIGINT
DECIMAL(p,s)	DECIMAL(p,s)
SMALLDECIMAL, CLOB, NCLOB, TEXT, TIME, SECONDDATE	STRING
REAL	FLOAT
DOUBLE	DOUBLE
BOOLEAN	BOOLEAN
VARCHAR(n), NVARCHAR(n)	VARCHAR(n)
CHAR(n), NCHAR(n)	CHAR(n)
DATE	DATE
TIMESTAMP	TIMESTAMP_NTZ
VARBINARY, BLOB	BINARY

CLI Reference

Databricks Publishing Arguments

Option	Type	Description
--publish_target ID	String	Credential ID for Databricks publishing (required)
--publish_schema_pattern PATTERN	String	Schema naming pattern (default: EXT_{schema} for external, {schema} for internal)
--publish_table_pattern PATTERN	String	Table naming pattern (default: {table})
--publish_method {external,internal}	Enum	external (Unity Catalog external table) or internal (managed Delta table) (default: external)
--pk_constraints	Flag	Add PRIMARY KEY constraints (managed/Delta tables only)
--n_jobs N	Integer	Parallel workers for table creation (default: 1)

There is no --databricks_catalog flag. The target catalog is always taken from the catalog field of the Databricks credential in credentials.json.

Deprecated / Renamed Arguments

Old Argument	Current Argument	Notes
--databricks_table_type	--publish_method {external|internal}	Renamed to the generic, cross-platform flag (internal maps to managed for Databricks)
--databricks_catalog	(none)	Never implemented as a CLI flag; the catalog comes from credentials only

Querying Tables

Databricks SQL:

SELECT * FROM lakexpress_catalog.tpch_1.customer LIMIT 10;

Databricks Notebooks (Python):

df = spark.table("lakexpress_catalog.tpch_1.customer")
df.display()

Databricks Notebooks (SQL):

%sql
SELECT * FROM lakexpress_catalog.tpch_1.customer LIMIT 10

Delta Table Features (Managed Tables Only)

Time Travel:

SELECT * FROM lakexpress_catalog.tpch_1.customer VERSION AS OF 1;

Table History:

DESCRIBE HISTORY lakexpress_catalog.tpch_1.customer;

Optimize:

OPTIMIZE lakexpress_catalog.tpch_1.customer;

See Also

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