的确是我错了。
- 我自己跑了个四万变量的,可以跑,就是双击打开有点那个......
- SAS HELP中What's New in the Base SAS 9.0, 9.1, and 9.1.3 Language有:The default BASE engine in SAS supports longer format and informat names, thread-enabled procedures such as the SORT and SUMMARY procedures, and more than 32,767 variables in a SAS data set.
- The Little SAS Book 中有:Prior to SAS 9.1, SAS data sets could contain up to 32,767 variables.
Beginning with SAS 9.1, the maximum number of variables in a SAS data set is limited by the
resources available on your computer,but SAS data sets with more than 32,767 variables cannot
be used with earlier versions of SAS. - mysas上看了一个问题"假如我们面对100万列的数据该怎么聚类?",oloolo大大的一个回答是:"SAS 9.1以上支持超过32K的变量个数,但是依然受pagesize和操作系统的限制,不是想多大就多大,你列的这个级别在PC上没法做,至少是32位windows下不行。"(链接http://www.mysas.net/forum/viewtopic.php?f=4&t=6114&start=10)。
以上几点至少说明SAS9.1(9.0?这个上2和3貌似有点小矛盾,还是偶的英语问题......,不管这些取个max总是对的)以后,SAS突破了2^15-1的变量限制......
ps,多想了下,oloolo大大说的原始数据(原作者假设为
1千万行×500万列)无法在32位windows下作,是不是由于32位文件系统的单个文件大小限制,偶估摸了下,那个数据貌似要10T级别......望高人指点