Glue Schema Registry to Iceberg schema type mapping

When you register a schema in the AWS Glue Schema Registry, the Channel maps the JSON Schema types to Iceberg column types as follows.

JSON Schema type	Condition	Iceberg column type
string	(plain)	string
string	format: "date-time"	timestamptz
string	format: "date"	date
string	format: "time"	time
string	format: "uuid"	uuid
string	format: "byte" or contentEncoding: "base64"	binary
integer	fits in 32-bit range	int
integer	exceeds 32-bit range	long
number	with multipleOf (for example, 0.001)	decimal(38, scale)
number	plain floating point	double
boolean	—	boolean
object	with named properties	struct
object	with additionalProperties	map<string, value_type>
array	—	list<element_type>
enum	—	string

Required columns — A column is marked required (non-nullable) when it appears in the "required" array of the parent object in your JSON Schema.

Partition key column — Your table must include a timestamptz column that can be used for time-based partitioning (hour). The source column referenced by the partition is automatically treated as required, even if your schema does not list it in the "required" array. You can enable an S3 Tables record-expiration job based on the table's partition column — see Managing S3 Tables record expiration.

The Channel handles source fields as follows:

• Extra fields in source data — Fields present in your source records but not defined in the table schema are silently dropped; they are not written to the Iceberg table.

• Missing optional fields — If a source record omits a field defined as optional in the table schema, the column is written as null.

• Missing required fields — If a source record omits a field defined as required, the record fails validation and is sent to the dead-letter queue (DLQ) with an error indicating the missing required field.

• Nesting limit — Schemas with more than 16 levels of nesting are not supported.