更新时间:2026-11-13 GMT+08:00
分享

Hive默认函数

Hive中有很多默认的函数,需要使用LOAD MODULE将Hive函数加载进来。加载命令如下:

LOAD MODULE hive WITH('hive-version' = '3.1.0');
USE MODULES `hive`, `core`;

注意事项

  • hive-version当前只支持3.1.0。
  • 如果配置了USE MODULES `hive`, `core`,则会优先使用Hive默认函数。

Hive默认函数使用示例

CREATE CATALOG lf_hive_catalog WITH (
  'type' = 'hive',
  'metastore'='lakeformation'
);

CREATE DATABASE IF NOT EXISTS lf_hive_catalog.hive_db
WITH (
  'location' = 'obs://fabricsql-hlt-test/aura-job/hive_db'
);

LOAD MODULE hive WITH('hive-version' = '3.1.0');
USE MODULES `hive`, `core`;

create table if not exists lf_hive_catalog.hive_db.hive_table (
  user_id STRING,
  item_id STRING,
  cat_id STRING,
  ts timestamp(9),
  dy STRING,
  ho STRING,
  mi STRING,
  `value` int
) PARTITIONED BY (dy, ho, mi)with
  (
    'connector' = 'hive',
    'partition.time-extractor.timestamp-pattern' = '$dy $ho:$mi:00',
    'sink.partition-commit.trigger' = 'process-time',
    'sink.partition-commit.delay' = '0S',
    'sink.partition-commit.policy.kind' = 'metastore,success-file',
    'hive.storage.file-format'='avro'
  );

CREATE TABLE datagen (
  user_id STRING,
  item_id STRING,
  cat_id STRING,
  ts timestamp(9),
  dy STRING,
  ho STRING,
  mi STRING
)
WITH
  (
    'connector' = 'datagen',
    'number-of-rows' = '10'
  );


insert into lf_hive_catalog.hive_db.hive_table
select user_id,item_id,cat_id,ts,dy,ho,mi,year('2026-10-13 12:08:56') from datagen;

Hive UDF与Hive默认函数混合使用样例

LOAD MODULE hive WITH('hive-version' = '3.1.0');
USE MODULES `hive`, `core`;

CREATE CATALOG lf_hive_catalogWITH (
  'type' = 'hive',
  'metastore'='lakeformation'
);

CREATE DATABASE IF NOT EXISTS lf_hive_catalog.hive_db
WITH (
  'location' = 'obs://fabricsql-hlt-test/aura-job/hive_db'
);

CREATE FUNCTION IF NOT EXISTS lf_hive_catalog.hive_db.hive_function AS 'org.AddDoublesUDF' USING JAR 'obs://bucket/SimpleUdf.jar';

create table if not exists lf_hive_catalog.hive_db.hive_table (
  user_id STRING,
  item_id STRING,
  cat_id STRING,
  ts timestamp(9),
  dy STRING,
  ho STRING,
  mi STRING,
  `year_value` int,
  `value` double
) PARTITIONED BY (dy, ho, mi)with
  (
    'connector' = 'hive',
    'partition.time-extractor.timestamp-pattern' = '$dy $ho:$mi:00',
    'sink.partition-commit.trigger' = 'process-time',
    'sink.partition-commit.delay' = '0S',
    'sink.partition-commit.policy.kind' = 'metastore,success-file',
    'hive.storage.file-format'='avro'
  );

CREATE TABLE datagen (
  user_id STRING,
  item_id STRING,
  cat_id STRING,
  ts timestamp(9),
  dy STRING,
  ho STRING,
  mi STRING,
  v_value DOUBLE,
  x_value DOUBLE
)
WITH
  (
    'connector' = 'datagen',
    'number-of-rows' = '10'
  );

insert into lf_hive_catalog.hive_db.hive_table
select user_id,item_id,cat_id,ts,dy,ho,mi,year('2026-10-13 12:08:56'),lf_hive_catalog.hive_db.hive_function(v_value,x_value) from datagen;

相关文档