更新时间:2026-11-13 GMT+08:00
Hive UDF注册到LakeFormation
注意事项
Hive UDF需要使用Hive Catalog来注册。注册和调用时需使用catalog_name.database_name.udf_name三段式命名语法。
Hive UDF准备
Hive UDF需要基于Hive接口实现,Hive Function实现后需要将其将打包后上传到OBS(上传操作详情请参见上传对象)。Hive UDF代码示例:
package org;
import org.apache.hadoop.hive.ql.exec.UDF;
public class AddDoublesUDF extends UDF {
public Double evaluate(Double... a) {
Double total = 0.0;
// 处理逻辑部分.
for (int i = 0; i < a.length; i++)
if (a[i] != null)
total += a[i];
return total;
}
} Hive UDF注册到LakeFormation
Hive UDF注册语法:
CREATE FUNCTION IF NOT EXISTS lf_hive_catalog.hive_db.hive_function AS 'class_name' USING JAR 'obs://bucket/SimpleUdf.jar';
注册的UDF可以在LakeFormation页面进行查看。

完整的使用示例
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'
);
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,
`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,lf_hive_catalog.hive_db.hive_function(v_value,x_value) from datagen; 父主题: Hive连接器