hive中使用sum能很方便地实现数值聚类求和,但如果对象是字符串,要求把字符串聚合成一个长串(用逗号分隔),最好还要去重,该如何实现?
下面的sql语句实现了在一个id对应多个图片(pic_id)的情况下,获得每个id对应的所有pic_id这一需求
select id,concat_ws(',',collect_set(pic_id)) as pic_ids from (
select id,pic_id from id_pic_mp
)fpg group by id
hive中使用sum能很方便地实现数值聚类求和,但如果对象是字符串,要求把字符串聚合成一个长串(用逗号分隔),最好还要去重,该如何实现?
下面的sql语句实现了在一个id对应多个图片(pic_id)的情况下,获得每个id对应的所有pic_id这一需求
select id,concat_ws(',',collect_set(pic_id)) as pic_ids from (
select id,pic_id from id_pic_mp
)fpg group by id