Question

我有一个表t1，有4列：

key, cd, date, result_num

在SAS中，我们有以下代码：

PROC SQL;
    create table t2 AS
    select * from t1
    group by key
    having date = MAX(date)
    order by key, cd;
RUN;

我的印象是，使用聚合函数（例如MAX）时所选的所有列都必须在group by中，或者应用了聚合函数。我的目标是将此SAS代码转换为SQL，有没有办法在SQL中执行此操作（更具体地说是hiveQL）？

Answer 1

我不认为您的查询在SAS中正在执行您想要的操作。。。也许确实如此。在标准SQL（和Hive）中，您可以执行以下操作：

create table t2 AS
    select *
    from (select t1.*,
                 row_number() over (partition by key order by date desc) as seqnum
          from t1
         ) t1
    where seqnum = 1
    order by key, cd;

Answer 2

诀窍是访问输入表两次：一次计算最大日期，一次选择适当的数据

如果您查找日期是整个表格中出现的最高日期的行，即

PROC SQL;
    create table t2 AS
    select * from t1
    where date = (select MAX(date) from t1)
    order by key, cd;
RUN;

如果您查找的日期是相同密钥的最高日期，即

PROC SQL;
    create table t2 AS
    select * from t1 inner join 
    (  select MAX(date) as maxDate 
       from t1  
       group by key) as m1 
       on m1.key = t1.key and m1.maxDate = t1.date
    order by key, cd;
RUN;

使用SQL查找具有极值的行

2 个答案: