我们常遇到一个存储过程要做都种排序的情况。通常order by 的条件只有一个固定的，这是远远不够的。需要用一个变量来替换它，而oracle有不能识别order by后面的字符串。该怎么办呢？先看看我的一个糟糕方案：
if myorderby=objectid desc then –按名称排序
open outcur for select * from(
select rownum rowno,t.* from (
select objectid , hits,posttime from cartoon order by objectid desc) t
where rownum=startrow;
elsif myorderby=hits desc then –按点击率排序
open outcur for select * from(
select rownum rowno,t.* from (
select objectid , hits,posttime from cartoon order by hits desc) t
where rownum=startrow;
else –按时间排序
open outcur for select * from(
select rownum rowno,t.* from (
select objectid , hits,posttime from cartoon order by hits desc) t
where rownum=startrow;
end if;
这虽然是通常使用的两种排序方式得以功过，但显得繁杂而且不够通用。
这时我的同事的好办法：
—————————————————————————
if (myorderby is not null) then
mysql:=select * from( select rownum rowno,t.* from (
select objectid , hits,posttimefrom cartoon order by || myorderby ||
) t where rownum< || endrow || ) where rowno>= || startrow;
else
mysql:=select * from( select rownum rowno,t.* from (
select objectid , hits,posttime from cartoon order by posttime desc) t
where rownum< || endrow || ) where rowno>= || startrow;
end if;
open outcur for mysql;
怎么样，感觉不同凡响吧。希望你能有更好的办法。
