一、TABLE METADATA
支持四段式的 Iceberg 表元数据查询,包括:history、snapshots、files、manifests。
二、语法
SELECT select_expr (, select_expr)*FROM `Catalog`.`db`.`tableName${history|snapshots|files|manifests|partitions|all_data_files|all_manifests}`[WHERE where_condition][LIMIT [offset,] rows]
三、示例
SELECT * FROM `DataLakeCatalog`.`validation`.`dempts$history` ORDER BY snapshot_id DESC LIMIT 1;SELECT * FROM `DataLakeCatalog`.`validation`.`dempts$snapshots` ORDER BY snapshot_id LIMIT 1;SELECT * FROM `DataLakeCatalog`.`validation`.`dempts$files` ORDER BY file_size_in_bytes LIMIT 1;SELECT * FROM `DataLakeCatalog`.`validation`.`dempts$manifests` ORDER BY length LIMIT 1;SELECT * FROM `DataLakeCatalog`.`validation`.`dempts$partitions` LIMIT 10;SELECT * FROM `DataLakeCatalog`.`validation`.`dempts$all_data_files` LIMIT 10;SELECT * FROM `DataLakeCatalog`.`validation`.`dempts$all_manifests` LIMIT 10;
四、FOR SYSTEM_TIME AS OF/ TIMESTAMP AS OF
Spark3.3以上支持,支持字符串和 Unix 时间戳两种格式
五、示例
SELECT empno FROM sales.emp FOR SYSTEM_TIME AS OF '1986-10-26 01:21:00';SELECT empno FROM sales.emp FOR SYSTEM_TIME AS OF 12324235546;SELECT empno FROM sales.emp TIMESTAMP AS OF '1986-10-26 01:21:00';SELECT empno FROM sales.emp TIMESTAMP AS OF 11111;
六、FOR SYSTEM_VERSION AS OF/ VERSION AS OF (Spark 3.3支持)
支持字符串和 snapshot id
七、示例
SELECT empno FROM sales.emp VERSION AS OF 'Snapshot123456789';SELECT empno FROM sales.emp VERSION AS OF 11111;SELECT empno FROM sales.emp FOR SYSTEM_VERSION AS OF 'Snapshot123456789';SELECT empno FROM sales.emp FOR SYSTEM_VERSION AS OF 11111;
八、示例
SELECT * FROM `sales`.`emp$at_timestamp_1111`;
九、示例
SELECT * FROM `sales`.`emp$snapshot_id_1111`