DDL 语法

最近更新时间:2023-07-19 18:06:52

我的收藏
说明:
以下语法说明为 DLC 原生表语法,DLC 原生表默认为 Iceberg 表。如果您使用 Iceberg 外部表,在 DDL 语法上会有细节差异,可参考文档 Iceberg 外部表与原生表语法差异

CREATE TABLE

语法

CREATE TABLE [ IF NOT EXISTS ] table_identifier
( col_name[:] col_type [ COMMENT col_comment ], ... )
[ COMMENT table_comment ]
[ PARTITIONED BY ( col_name1, transform(col_name2), ... ) ]

参数

table_identifier

table_identifier 支持三段式:catalog.db.name

Schemas and Data Types

col_type
: primitive_type
  | nested_type
primitive_type
: boolean
| int/integer
| long/bigint
| float
| double
| decimal(p,s),p=最大位数,s=最大小数点位数, s<=p<=38
| date
| timestamptimestamp with timezone,不支持time和without timezone
| string,也可对应Iceberg uuid类型
| binary,也可对应Iceberg fixed类型
nested_type
: struct
| list
| map

Partition Transforms

transform
: identity,支持任意类型, DLC不支持该转换
| bucket[N]hash mod N分桶,支持col_type: int,long, decimal, date, timestamp, string, binary
| truncate[L],L截取分桶,支持col_type: int,long,decimal,string
| years,年份,支持col_type: date,timestamp
| months,月份,支持col_type: date,timestamp
| days/date,日期,支持col_type: date,timestamp
| hours/date_hour,小时,支持col_type: timestamp

示例

CREATE TABLE dempts(
id bigint COMMENT 'id number',
num int,
eno float,
dno double,
cno decimal(9,3),
flag  boolean,
data string,
ts_year timestamp,
date_month date,
bno binary,
point struct<x: double, y: double>,
points array<struct<x: double, y: double>>,
pointmaps map<struct<x: int>, struct<a: int>>
)
COMMENT 'table documentation'
PARTITIONED BY (bucket(16,id), years(ts_year), months(date_month), identity(bno), bucket(3,num),truncate(10,data));

CREATE TABLE AS SELECT

语法

CREATE TABLE [ IF NOT EXISTS ] table_identifier
[ COMMENT table_comment ]
[ PARTITIONED BY ( col_name1, transform(col_name2), ... ) ]
[ TBLPROPERTIES ( property_name=property_value, ... ) ]
AS select_statement

示例

CREATE TABLE dempts_copy
COMMENT 'table create as select'
PARTITIONED BY (eno, dno)
AS SELECT * from dempts;

REPLACE TABLE AS SELECT

保留表历史 History 的情况下,用 SELECT 查询的结果生成一个快照 Snapshot 来更新表。

语法

CREATE [OR REPLACE] TABLE table_identifier
[ COMMENT table_comment ]
[ PARTITIONED BY ( col_name1, transform(col_name2), ... ) ]
AS select_statement

示例

CREATE OR REPLACE TABLE dempts_replace
COMMENT 'table create as replace'
PARTITIONED BY (eno, dno)
AS SELECT * from dempts;

DROP TABLE

语法

DROP TABLE [ IF EXISTS ] table_identifier

ALTER TABLE

变更表语法

ALTER TABLE ... RENAME TO

变更表名称

语法

ALTER TABLE table_identifier RENAME TO new_table_identifier

ALTER TABLE ... SET / UNSET TBLPROPERTIES 

更新/删除数据表属性

语法

-- SET 更新属性配置
ALTER TABLE table_identifier
SET TBLPROPERTIES (property_name=property_value, ...)

-- UNSET 删除属性配置
ALTER TABLE table_identifier
UNSET TBLPROPERTIES (property_name, ...)

示例

-- SET 更新属性配置
ALTER TABLE dempts SET TBLPROPERTIES ('read.split.target-size'='268435456');

-- UNSET 删除属性配置
ALTER TABLE dempts SET TBLPROPERTIES ('read.split.target-size'='268435456');


ALTER TABLE ... WRITE ORDERED BY

设置表数据插入时的排序方式

语法

ALTER TABLE table_identifier 
WRITE [LOCALLY] ORDERED BY 
{col_name [ASC|DESC] [NULLS FIRST|LAST]}[, ...]

示例

ALTER TABLE dempts WRITE ORDERED BY category, id;

-- use optional ASC/DEC keyword to specify sort order of each field (default ASC)
ALTER TABLE dempts WRITE ORDERED BY category ASC, id DESC;

-- use optional NULLS FIRST/NULLS LAST keyword to specify null order of each field (default FIRST)
ALTER TABLE dempts WRITE ORDERED BY category ASC NULLS LAST, id DESC NULLS FIRST;

-- To order within each task, not across tasks
ALTER TABLE dempts WRITE LOCALLY ORDERED BY category, id;

ALTER TABLE ... WRITE DISTRIBUTED BY PARTITION

修改分区表的数据分配策略

语法

ALTER TABLE table_identifier 
WRITE DISTRIBUTED BY PARTITION 
[ LOCALLY ORDERED BY 
{col_name [ASC|DESC] [NULLS FIRST|LAST]}[, ...]] 

示例

ALTER TABLE dempts WRITE DISTRIBUTED BY PARTITION;
ALTER TABLE dempts WRITE DISTRIBUTED BY PARTITION LOCALLY ORDERED BY id;

ALTER TABLE COLUMNS

变更字段语法

ALTER TABLE ... ADD COLUMNS

新增多个字段

语法

-- 新增多个字段
ALTER TABLE table_identifier 
ADD COLUMNS (col_name col_type [COMMENT col_comment], ...)

-- 新增单个字段
ALTER TABLE table_identifier 
ADD COLUMN col_name col_type [COMMENT col_comment] 
[FIRST | AFTER target_col_name]

示例

-- 新增多个字段
ALTER TABLE dempts
ADD COLUMNS (
new_column_1 string comment 'new_column_1 docs',
new_column_2 int comment 'new_column_2 docs'
);

-- 新增单个字段
ALTER TABLE dempts 
ADD COLUMN new_column_3 string comment 'new_column docs';

ALTER TABLE ... RENAME COLUMN 

变更字段名称

语法

ALTER TABLE table_identifier 
RENAME COLUMN old_column_name TO new_column_name

ALTER TABLE ... ALTER COLUMN

变更字段类型/备注信息

语法

ALTER TABLE table_identifier 
ALTER COLUMN col_name 
{TYPE new_col_type | COMMENT col_comment}
目前Iceberg TYPE 变更,仅支持字段类型安全扩展:
int/integer -> long/bigint
float -> double
decimal(P,S) -> decimal(P2,S),其中P2 > P,即精准度提升

示例

ALTER TABLE dempts ALTER COLUMN new_column_2 TYPE bigint;
ALTER TABLE dempts ALTER COLUMN new_column_2 comment 'alter docs';

ALTER TABLE ... DROP COLUMN

删除表字段

语法

ALTER TABLE table_identifier DROP COLUMN column_name

ALTER TABLE PARTITIONS

ALTER TABLE ... ADD PARTITION FIELD

新增单个分区字段

语法

ALTER TABLE table_identifier 
ADD PARTITION FIELD col_name|transform (col_name) [AS alias]
transform 参考 CREATE TABLE 说明。

示例

ALTER TABLE dempts ADD PARTITION FIELD new_column_1;
ALTER TABLE dempts ADD PARTITION FIELD bucket(3,new_column_2);

ALTER TABLE ... REPLACE PARTITION FIELD

替换单个分区字段

语法

ALTER TABLE table_identifier REPLACE PARTITION FIELD col_name|transform (col_name) [AS alias]

ALTER TABLE ... DROP PARTITION FIELD

删除单个分区字段

语法

ALTER TABLE table_identifier 
DROP PARTITION FIELD col_name|transform (col_name);
transform 参考 CREATE TABLE 说明

示例

ALTER TABLE dempts DROP PARTITION FIELD new_column_1;
ALTER TABLE dempts DROP PARTITION FIELD bucket(3,new_column_2);

ALTER TABLE ... SET IDENTIFIER FIELDS 

添加 identifier fields 属性

语法

ALTER TABLE dempts SET IDENTIFIER FIELDS empno, name

ALTER TABLE .. DROP IDENTIFIER FIELDS

添加 identifier fields 属性

语法


ALTER TABLE dempts DROP IDENTIFIER FIELDS empno, name