Implementation of core data structures for R.
Implementation of advanced data structures such as hashmaps, heaps, or queues in R
.
Advanced data structures are essential in many computer science and statistics
problems, for example graph algorithms or string analysis. The package uses
Boost
and STL
data types and extends these to R
with Rcpp
modules.
So far datastructures
has implementations for:
- Fibonacci and binomial heaps,
- queues and stacks,
- hashmaps and bimaps.
As an introductory example, consider you want to compute shortest paths on a graph and decide to use a Fibonacci heap for keeping the distances. A Fibonacci heap is an efficient tree-like data structure that satisfies the min-heap property. We can use it to quickly get the node with the shortest distance in O(log n) time like this:
fh <- fibonacci_heap("character", "numeric")
node.labels <- paste0("n", 10:1)
node.distances <- seq(1, 0, length.out=length(node.labels))
fh <- insert(fh, node.labels, node.distances)
peek(fh)
$n1
[1] 0
Get the package from CRAN using:
install.packages("datastructures")
You can also download the tarball of the latest release and install with:
R CMD install <datastructures-x.y.z.tar.gz>
where <datastructures-x.y.z.tar.gz>
is your downloaded tarball. If you want
to you can also use devtools, but I don't recommend it since it might give unstable
versions:
devtools::install_github("dirmeier/datastructures")
Load the library using library(datastructures)
. We provide a vignette for
the package that can be called using: vignette("datastructures")
. If there
are any questions let met know.
If you want to have another datastructure added, say from boost
or the STL
,
just open up a new issue. Alternatively it would be great if you provided a PR.
- Simon Dirmeier [email protected]