hive中的concat，concat_ws，collect_set用法

需求：对用户的信息进行分析，相同用户的地址信息按照不同类型分别展示出来，每个用户一行。

table1:

user_id	location	location_type
123	w2ny6s	home
123	w2ny6k	work

目标：

user_id	home_location	work_location
123	w2ny6s	w2ny6k

使用concat_ws()和collect_set()进行合并行，将上面列表中一个user可能会占用多行转换为每个user占一行的目标表格式，实际是“列转行”

    select user_id,
           concat_ws('',collect_set(home_location)) as home_location,
           concat_ws('',collect_set(work_location)) as work_location
      from (
            select user_id,
                   if(location_type='home', location, "") as home_location,
                   if(location_type='work', location, "") as work_location
              from table1
           ) a

collect_set的作用：collect_set(col)函数只接受基本数据类型，它的主要作用是将某字段的值进行去重汇总，产生array类型字段。

concat_ws的作用：表示concat with separator,即有分隔符的字符串连接，concat_ws(”,collect_set(home_location))表示用空的字符”来连接collect_set返回的array中的每个元素。

concat：可以连接一个或者多个字符串，select concat(‘11’,’22’,’33’);//112233