Sort Buffers
Documentation | Blog | Demos | Support
Sort Buffers
5 out of 5 stars
1 rating
5 Stars | 100% | |
4 Stars | 0% | |
3 Stars | 0% | |
2 Stars | 0% | |
1 Stars | 0% |
sort buffers nn
where nn is the maximum amount of memory (KB) that can be used when sorting data. If the requested amount of memory is not available, all available memory up to the limit specified by nn is used.
During a sorting process, fields of data to be sorted are extracted and put into sort records that are placed into sort buffers. Then, either the entire sort is performed in available memory or, if not enough memory is available, sort records spill into temporary disk files (the zimstxxx files).
A sort that occurs entirely in memory is quickest. If a sort process must use disk space, the records placed in temporary disk files must be merged back into the sorted order, resulting in reduced performance.
A high value for sort buffers has two effects:
- The sorting is able to take place completely in memory.
- The merging process is faster, as fewer passes are required to merge the data in temporary disk files.
Sort buffers space is acquired only during a sorting process and is returned once the sort is completed.
Set operations (UNION, INTERSECT, and MINUS) and the KEEP subcommand generate sort operations internally. Their speed is therefore also affected by the amount of sort buffer space.
Valid Settings
nn = numeric entry 100 to 10000
Windows | 1000 |
UNIX | 1000 |
5 out of 5 stars
1 rating
5 Stars | 100% | |
4 Stars | 0% | |
3 Stars | 0% | |
2 Stars | 0% | |
1 Stars | 0% |