relate {wakefield}

Create Related Numeric Columns

Description

Generate columns that are related.

Usage

relate(
  x,
  j,
  name = NULL,
  operation = "+",
  mean = 5,
  sd = 1,
  rep.sep = "_",
  digits = max(nchar(sub("^[^.]*.", "", x)))
)

Arguments

x

A starting column.

j

The number of columns to produce.

name

An optional prefix name to give to the columns. If NULL attempts to take from the varname attribute of x. If not found, "Variable" is used.

operation

A operation character vector of length 1; either c("+", "-", "*", "/"). This is the relationship between columns.

mean

Mean is the average value to add, subtract, multiple, or divide by.

sd

The amount of variability to allow in mean. Setting to 0 will constrain the operation between x_(n - 1) column and x_n to be exactly the mean value (see Examples for a demonstration).

rep.sep

A separator to use for repeated variable names. For example if the age is used three times (r_data_frame(age, age, age)), the name "Age" will be assigned to all three columns. The results in column names c("Age_1", "Age_2", "Age_3").

digits

The number of digits to round to. Defaults to the max number of significant digits in x.

Value

Returns a tbl_df.

See Also

r_series

Examples

relate(1:10, 10)

(x <- r_data_frame(10, id, relate(1:10, 10, "Time", mean = 2)))
library(ggplot2)

dat <- with(x, data.frame(ID = rep(ID, ncol(x[, -1])), stack(x[, -1])))
dat[["Time"]] <- factor(sub("Time_", "", dat[["ind"]]), levels = 1:10)
ggplot(dat, aes(x = Time, y = values, color = ID, group = ID)) +
    geom_line(size=.8)

relate(1:10, 10, name = "X", operation = "-")
relate(1:10, 10, "X", mean = 1, sd = 0)
relate(1:10, 10, "Var", "*")
relate(1:10, 10, "Var", "/")

relate(gpa(30), 5, mean = .1)
relate(likert(10), 5, mean = .1, sd = .2)
relate(date_stamp(10), 6)
relate(time_stamp(10), 6)
relate(rep(100, 10), 6, "Reaction", "-")