IndxVec

Usage

Insert into your Cargo.toml file [dependencies] section:

indxvec = "^0.1"

Import into your source file(s) macro here, struct GS, functions and trait as you need. There is just one trait Indices implemented on indices of type &[usize]. There is a bunch of functions in module merge which usually take generic vector(s) as arguments and may produce some indices.

use indxvec::{here,GS,merge::*,Indices};

See tests/tests.rs for examples of usage. To run the tests, use single thread. It may be slower but will produce the results in the right order:

cargo test --release -- --test-threads=1 --nocapture --color always

Description

Indxvec is a spin-off from rstats. It is a self-contained unit, both in terms of the subject matter and also in not having any dependencies at all.

The tools included are: efficient ranking, sorting, merging, searching and indices manipulations. They are applicable to generic vectors Vec<T> (or generic slices &[T]), thus they will work on primitive types but also on any arbitrarily complex end type T, as long as you implement their required traits, mostly just PartialOrd and/or Copy for T.

Functions

are in the module src/merge.rs. They mostly take some generic data and produce the indices onto which the methods of the following trait can be conveniently chained. See the documentation.

Trait Index

The methods of this trait are implemented for vectors of subscripts, i.e. &[usize].

invindex - method for inverting an index, e.g. given a sort index, returns ranks and vice versa.
unindex - collects values from a vector in the order given by an index. This will, for example, sort a vector into sort order when supplied with a sort index. Can also be used for key sort, whereby a vector of keys is extracted from some data, then a sort index is obtained for the keys (with sortidx) and that is applied (with unindex) to the original data. Also, the original sort order can be easily and efficiently reversed by supplying a boolean false second argument to unindex. This without repeating the sort or reversing the potentially bulky sorted data.
ucorrelation - Pearson's correlation coefficient of two indices, typically ranks. This is the same as Spearman's correlation of the original data.

Release Notes (Latest First)

Version 0.1.7 - added convenience conversion method indx_to_f64.

Version 0.1.6 - improved comments. Used Vec::with_capacity for new vectors of known lengths. Maybe a bit faster but no change in functionality.

Version 0.1.5 - fixed an inconsistency in binsearch result.

Version 0.1.4 - swapped arguments of unindex for compatibility. Added more comments.