stringdist R Package

You can install the stable version of stringdist from CRAN using the standard R installation command.

If you need the absolute latest development version from source, you can clone the GitHub repository and build it manually. Warning: The development version is written in C and may be unstable or cause R sessions to crash.

install.packages('stringdist')

Developers writing C or C++ code in other packages can call stringdist's underlying C functions directly. To access the documentation for the C API, you can either type ?stringdist_api in the R console or open the specific vignette using the command below.

vignette("stringdist_C-Cpp_api", package="stringdist")

To build the package from the latest source code, use a bash terminal (or Git Bash on Windows) to clone, build, and install the resulting tarball.

git clone https://github.com/markvanderloo/stringdist.git
cd stringdist
bash ./build.bash
R CMD INSTALL output/stringdist_*.tar.gz

As of version 0.9.5.0, several arguments that were previously used for parallelization or distance limiting have been removed. If you are upgrading from a very old version, ensure you remove the following from your function calls:

• cluster (for stringdistmatrix)
• maxDist (for stringdist and stringdistmatrix, but NOT amatch)
• ncores (for stringdistmatrix)

Parallelization Note: The package now uses openMP for parallel computing instead of R's parallel package, making it more efficient and system-independent.

The stringdist package provides several primary functions for approximate string matching and distance calculations in R:

• stringdist: Computes pairwise distances between two input character vectors (the shorter vector is recycled).
• stringdistmatrix: Computes the distance matrix for one or two vectors.
• stringsim: Computes a string similarity score between 0 and 1, based on stringdist.
• amatch: A fuzzy matching equivalent of R's native match function.
• ain: A fuzzy matching equivalent of R's native %in% operator.
• afind: Finds the location of fuzzy matches of a short string within a long string.
• seq_dist, seq_distmatrix, seq_amatch, and seq_ain: Specialized versions for computing distances and matching for integer sequences.

For tasks requiring fuzzy matching (finding approximate matches rather than exact ones), use these functions which act as equivalents to native R functions:

• amatch: A fuzzy matching equivalent of R's native match function.
• ain: A fuzzy matching equivalent of R's native %in% operator.

If you are working with integer sequences rather than character strings, use the following specialized functions:

• seq_dist: Distance between integer sequences.
• seq_distmatrix: Distance matrix for integer sequences.
• seq_amatch: Fuzzy matching for integer sequences.
• seq_ain: Fuzzy %in% equivalent for integer sequences.

Since version 0.9.5.0, you can call stringdist functions directly from the C code of your own R package. To access the full API documentation, you can:

1. Type ?stringdist_api in the R console.
2. Browse the package help index under User guides, package vignettes and other documentation and select doc/stringdist_api.pdf.
3. Locate the PDF directly using:

system.file("doc/stringdist_api.pdf", package="stringdist")

Use these utility functions for phonetic encoding and character analysis:

• qgrams(): Tabulates qgrams in one or more character vectors.
• seq_qrams(): Tabulates qgrams (ngrams) in one or more integer vectors.
• phonetic(): Computes phonetic codes (currently only supports Soundex).
• printable_ascii(): Detects non-printable ASCII or non-ASCII characters.

The package includes several utilities for string processing and analysis:

• qgrams(): Tabulates qgrams in one or more character vectors.
• seq_qrams(): Tabulates qgrams (sometimes called ngrams) in one or more integer vectors.
• phonetic(): Computes phonetic codes of strings (currently supports soundex).
• printable_ascii(): Detects non-printable ASCII or non-ASCII characters.

The package provides several core functions for calculating distances and similarities between character vectors:

• stringdist: Computes pairwise distances between two input character vectors (the shorter vector is recycled).
• stringdistmatrix: Computes a distance matrix for one or two vectors.
• stringsim: Computes a string similarity score between 0 and 1, derived from stringdist.