S******y
发帖数: 1123 | 1 What is the fastest way of look-up in SAS (say, matching between a small
file and a huge master-list file via id)? |
p********a
发帖数: 5352 | 2 I always use proc formats in this case |
l***a
发帖数: 12410 | 3 for easy coding and flexibility I use proc sql. if speed is a concern I will
do hashing.
index better or not depends on if id is heavily duplicated in the master
file. if each id is unique, index will on the contrary increase the
processing time due to more i/o
【在 S******y 的大作中提到】
: What is the fastest way of look-up in SAS (say, matching between a small
: file and a huge master-list file via id)?
|
s*r
发帖数: 2757 | 4 sas的hash做得很好吗
will
【在 l***a 的大作中提到】
: for easy coding and flexibility I use proc sql. if speed is a concern I will
: do hashing.
: index better or not depends on if id is heavily duplicated in the master
: file. if each id is unique, index will on the contrary increase the
: processing time due to more i/o
|
l***a
发帖数: 12410 | 5 compared to couple of proc sql and the saving on time is noticable, say, 50%
【在 s*r 的大作中提到】
: sas的hash做得很好吗
:
: will
|
o****o
发帖数: 8077 | 6 small look up in huge master table, I always bet on hash table, you will
then do a sequential read on the master which is the fastest way comparing
to other methods that involves non-sequential read over multiple tables
【在 S******y 的大作中提到】
: What is the fastest way of look-up in SAS (say, matching between a small
: file and a huge master-list file via id)?
|
r********e
发帖数: 1686 | 7 好无知啊，啥是SAS HASH？
555555555
50%
【在 l***a 的大作中提到】
: compared to couple of proc sql and the saving on time is noticable, say, 50%
|
s**********e
发帖数: 63 | |
