Use these simple techniques to improve SAS processing by minimize the I/O operations….
1.Use LENGTH statements to minimize variable lengths, where possible.
2. Use CLASS statement(s), where possible, rather than BY statements, which might require a SORT.
3. Use DROP and KEEP statements to minimize observation length.
4. Use SORT only when necessary.
5. Create multiple data sets in one DATA step when possible.
6. Use WHERE statements with procedures to avoid subsetting in a DATA step followed by a procedure.
7. Use the _NULL_ DATA step when you don’t need to create a data set. For example, you might want to create an external file, produce a report, or just do some calculations.
8. Compress some large SAS data sets, but beware that compression can use more space in some cases, which might actually increase I/O.
9. Develop and test programs on a small subset of the data.
2. Use CLASS statement(s), where possible, rather than BY statements, which might require a SORT.
3. Use DROP and KEEP statements to minimize observation length.
4. Use SORT only when necessary.
5. Create multiple data sets in one DATA step when possible.
6. Use WHERE statements with procedures to avoid subsetting in a DATA step followed by a procedure.
7. Use the _NULL_ DATA step when you don’t need to create a data set. For example, you might want to create an external file, produce a report, or just do some calculations.
8. Compress some large SAS data sets, but beware that compression can use more space in some cases, which might actually increase I/O.
9. Develop and test programs on a small subset of the data.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.