Download
headsort sorts files by line and fields. Options can pick field order and etc.
In general headsort can sort in 1/2 the time sort(1) (mergesort) does in situations involving either pipes or mult column selection; less than 1/2 with both. Best case was about opposite that of sort(1)*.
NEW* hsort main algorithm speed improvement (applies to all; 27,29,31). Now faster than sort(1) (in all cases tested).
Using triangulation, hsort begins piped sorted output when 1st bin is done. This means it has new applications in distributed computing and in some cases makes dc un-necessary. (gnu sort must finish all bins)
Well compatible with gnu sort (not perfect).
other algorthms featured if menu used:
bubble, selection, insertion, shell, distributive counting, straight radix, radix exchange, quick, merge. (some are by character, some by line, some both)
Features
- NEW: main algo SPEED improvement ! (all: 27,29,31)
- NEW: mini-sort DEMO for those wishing to use hsort algorithm in other programs
- exact column selection sort what's needed , not all / heuristic
- non-buggy numeric sorting , fast , very wide numbers
- floating month names (as international) (or as ordering rule)
- pretty text + binary + hex dump of file
- menu for trying other algorithms / selectable algorithm
- all sort(1) options supported , always stable
- arg handling even closer to gnusort's
- sources much cleaned and simplified. easy GNU make.
- see manpage for all features
- (new speed by dimensional access reduction, 31 now has two kinds of)
- 4.1.4 fixes minor file open issue (found needed when stdin is broken)
Project Samples
