depends on you data
if those [a,b] intervals are mutually exclusive, there're ways to fine-tune the join.
on the other hand, if there are many overlaps, it's beyond me.

Try if this code
The results may be slightly different from that of sql, due to the replicate.

谢谢。确实达到了目的，还有一个问题：hash里使用不等式条件是不是比使用等式条件慢？

"If neither contributing dataset is sorted, but one of them does fit into memory, PROC SQL will choose to process the query with a Hash Join (sqxjhsh)" (http://support.sas.com/techsup/technote/ts553.html)
Make sure you have enough memory, and the SQL join will be as efficient as Yongyitian's program.
The option is buffersize (http://support.sas.com/resources ... ings10/139-2010.pdf)
In either method, a Cartisian product is involved, so they should be spending very similar amount of time. It's not a problem as long as you can get your results in a reasonable amount of time.

"If neither contributing dataset is sorted, but one of them does fit into memory, PROC SQL will choose to process the query with a Hash Join (sqxjhsh)" (http://support.sas.com/techsup/technote/ts553.html)
Make sure you have enough memory, and the SQL join will be as efficient as Yongyitian's program.
The option is buffersize (http://support.sas.com/resources ... ings10/139-2010.pdf)
In either method, a Cartisian product is involved, so they should be spending very similar amount of time. It's not a problem as long as you can get your results in a reasonable amount of time.