radixorder: Fast Radix-Based Ordering

In SebKrantz/collapse: Advanced and Fast Data Transformation

Fast Radix-Based Ordering

Description

A slight modification of order(..., method = "radix") that is more programmer friendly and, importantly, provides features for ordered grouping of data (similar to data.table:::forderv which has more or less the same source code). radixorderv is a programmers version directly supporting vector and list input.

Usage

radixorder(..., na.last = TRUE, decreasing = FALSE, starts = FALSE,
           group.sizes = FALSE, sort = TRUE)

radixorderv(x, na.last = TRUE, decreasing = FALSE, starts = FALSE,
            group.sizes = FALSE, sort = TRUE)

Arguments

...	comma-separated atomic vectors to order.
x	an atomic vector or list of atomic vectors such as a data frame.
na.last	logical. for controlling the treatment of NA's. If TRUE, missing values in the data are put last; if FALSE, they are put first; if NA, they are removed.
decreasing	logical. Should the sort order be increasing or decreasing? Can be a vector of length equal to the number of arguments in ... / x.
starts	logical. TRUE returns an attribute 'starts' containing the first element of each new group i.e. the row denoting the start of each new group if the data were sorted using the computed ordering vector. See Examples.
group.sizes	logical. TRUE returns an attribute 'group.sizes' containing sizes of each group in the same order as groups are encountered if the data were sorted using the computed ordering vector. See Examples.
sort	logical. This argument only affects character vectors / columns passed. If FALSE, these are not ordered but simply grouped in the order of first appearance of unique elements. This provides a slight performance gain if only grouping but not alphabetic ordering is required. See also group.

Value

An integer ordering vector with attributes: Unless na.last = NA an attribute "sorted" indicating whether the input data was already sorted is attached. If starts = TRUE, "starts" giving a vector of group starts in the ordered data, and if group.sizes = TRUE, "group.sizes" giving the vector of group sizes are attached. In either case an attribute "maxgrpn" providing the size of the largest group is also attached.

Author(s)

The C code was taken - with slight modifications - from base R source code, and is originally due to data.table authors Matt Dowle and Arun Srinivasan.

See Also

Fast Grouping and Ordering, Collapse Overview

Examples

radixorder(mtcars$mpg)
head(mtcars[radixorder(mtcars$mpg), ])
radixorder(mtcars$cyl, mtcars$vs)

o <- radixorder(mtcars$cyl, mtcars$vs, starts = TRUE)
st <- attr(o, "starts")
head(mtcars[o, ])
mtcars[o[st], c("cyl", "vs")]  # Unique groups

# Note that if attr(o, "sorted") == TRUE, then all(o[st] == st)
radixorder(rep(1:3, each = 3), starts = TRUE)

# Group sizes
radixorder(mtcars$cyl, mtcars$vs, group.sizes = TRUE)

# Both
radixorder(mtcars$cyl, mtcars$vs, starts = TRUE, group.sizes = TRUE)

SebKrantz/collapse documentation built on Dec. 16, 2024, 7:26 p.m.