Fixes #350

NEWS.md

@@ -12,6 +12,9 @@ None
 
 ## Minor Fixes and Improvements
 1. Fix #372 and #373 - Update `cvar()` and `csd()` to a vectorized approach from @kokbent
+which speeds these up by over 100x
+2. Fix #350 - Update all `tidy_` distribution functions to generate data using `data.table` 
+this in many instances has resulted in a speed up of 30% or more.
 
 # TidyDensity 1.2.6
 

R/00_global_variables.R

@@ -7,6 +7,6 @@ globalVariables(
     "name", "prob", "rate", "sd_log", "shape1", "shape2", "size", "total", "total_deviance",
     "abs_aic", "aic_value", "data", "dist_with_params", "ks", "lm", "lm_model", "mu",
     "stan_dev", "tidy_ks", "prop", "dens_tbl", "cih", "cil", "mstat", "stat", ".",
-    "variable"
+    "variable",".N"
   )
 )

R/autoplot-multi-dist.R

@@ -45,7 +45,8 @@
 #'     .n = 500,
 #'     .mean = c(-2, 0, 2),
 #'     .sd = 1,
-#'     .num_sims = 5
+#'     .num_sims = 5,
+#'     .return_tibble = TRUE
 #'   )
 #' )
 #'

R/combine-multi-single-dist-tbl.R

@@ -24,7 +24,8 @@
 #'     .n = 50,
 #'     .mean = c(-1, 0, 1),
 #'     .sd = 1,
-#'     .num_sims = 3
+#'     .num_sims = 3,
+#'     .return_tibble = TRUE
 #'   )
 #' )
 #'

R/random-tidy-bernoulli.R

@@ -33,8 +33,8 @@
 #' @param .n The number of randomly generated points you want.
 #' @param .prob The probability of success/failure.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_bernoulli()

R/random-tidy-beta.R

@@ -34,11 +34,12 @@
 #' @param .shape2 A non-negative parameter of the Beta distribution.
 #' @param .ncp The `non-centrality parameter` of the Beta distribution.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_beta()
+#'
 #' @return
 #' A tibble of randomly generated data.
 #' @name tidy_beta

R/random-tidy-binomial.R

@@ -32,8 +32,8 @@
 #' @param .size Number of trials, zero or more.
 #' @param .prob Probability of success on each trial.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_binomial()

R/random-tidy-burr-inverse.R

@@ -35,8 +35,8 @@
 #' @param .scale Must be strictly positive.
 #' @param .rate An alternative way to specify the `.scale`.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_inverse_burr()

R/random-tidy-burr.R

@@ -34,8 +34,8 @@
 #' @param .scale Must be strictly positive.
 #' @param .rate An alternative way to specify the `.scale`.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_burr()

R/random-tidy-cauchy.R

@@ -32,8 +32,8 @@
 #' @param .location The location parameter.
 #' @param .scale The scale parameter, must be greater than or equal to 0.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_cauchy()

R/random-tidy-chisquare.R

@@ -32,8 +32,8 @@
 #' @param .df Degrees of freedom (non-negative but can be non-integer)
 #' @param .ncp Non-centrality parameter, must be non-negative.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_chisquare()
@@ -53,7 +53,7 @@ tidy_chisquare <- function(.n = 50, .df = 1, .ncp = 1, .num_sims = 1,
   # Tidyeval ----
   n <- as.integer(.n)
   num_sims <- as.integer(.num_sims)
-  df <- as.numeric(.df)
+  deg_f <- as.numeric(.df)
   ncp <- as.numeric(.ncp)
   ret_tbl <- as.logical(.return_tibble)
 
@@ -72,13 +72,13 @@ tidy_chisquare <- function(.n = 50, .df = 1, .ncp = 1, .num_sims = 1,
     )
   }
 
-  if (!is.numeric(df) | !is.numeric(ncp)) {
+  if (!is.numeric(deg_f) | !is.numeric(ncp)) {
     rlang::abort(
       "The parameters of .df and .ncp must be of class numeric."
     )
   }
 
-  if (df < 0 | ncp < 0) {
+  if (deg_f < 0 | ncp < 0) {
     rlang::abort("The parameters of .df and .ncp must be greater than or equal to 0.")
   }
 
@@ -92,16 +92,16 @@ tidy_chisquare <- function(.n = 50, .df = 1, .ncp = 1, .num_sims = 1,
   df <- data.table::CJ(sim_number = factor(1:num_sims), x = 1:n)
 
   # Group the data by sim_number and add columns for x and y
-  df[, y := stats::rchisq(n = .N, df = df, ncp = ncp)]
+  df[, y := stats::rchisq(n = .N, df = deg_f, ncp = ncp)]
 
   # Compute the density of the y values and add columns for dx and dy
   df[, c("dx", "dy") := density(y, n = n)[c("x", "y")], by = sim_number]
 
   # Compute the p-values for the y values and add a column for p
-  df[, p := stats::pchisq(y, df = df, ncp = ncp)]
+  df[, p := stats::pchisq(y, df = deg_f, ncp = ncp)]
 
   # Compute the q-values for the p-values and add a column for q
-  df[, q := stats::qchisq(p, df = df, ncp = ncp)]
+  df[, q := stats::qchisq(p, df = deg_f, ncp = ncp)]
 
   if(.return_tibble){
     df <- dplyr::as_tibble(df)

R/random-tidy-exponential-inverse.R

@@ -33,8 +33,8 @@
 #' @param .scale Must be strictly positive.
 #' @param .rate An alternative way to specify the `.scale`
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_inverse_exponential()

R/random-tidy-exponential.R

@@ -31,8 +31,8 @@
 #' @param .n The number of randomly generated points you want.
 #' @param .rate A vector of rates
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_exponential()

R/random-tidy-f.R

@@ -33,8 +33,8 @@
 #' @param .df2 Degrees of freedom, Inf is allowed.
 #' @param .ncp Non-centrality parameter.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_f()

R/random-tidy-gamma-inverse.R

@@ -34,8 +34,8 @@
 #' @param .scale Must be strictly positive.
 #' @param .rate An alternative way to specify the `.scale`
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_inverse_gamma()

R/random-tidy-gamma.R

@@ -34,8 +34,8 @@
 #' @param .scale The standard deviation of the randomly generated data. This is
 #' strictly from 0 to infinity.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_gamma()
@@ -97,16 +97,16 @@ tidy_gamma <- function(.n = 50, .shape = 1, .scale = 0.3, .num_sims = 1,
   df <- data.table::CJ(sim_number = factor(1:num_sims), x = 1:n)
 
   # Group the data by sim_number and add columns for x and y
-  df[, y := stats::rgamma(n = .N, shape = shape, scale = scle)]
+  df[, y := stats::rgamma(n = .N, shape = shp, scale = scle)]
 
   # Compute the density of the y values and add columns for dx and dy
   df[, c("dx", "dy") := density(y, n = n)[c("x", "y")], by = sim_number]
 
   # Compute the p-values for the y values and add a column for p
-  df[, p := stats::pgamma(y, shape = shape, scale = scle)]
+  df[, p := stats::pgamma(y, shape = shp, scale = scle)]
 
   # Compute the q-values for the p-values and add a column for q
-  df[, p := stats::qgamma(p, shape = shape, scale = scle)]
+  df[, p := stats::qgamma(p, shape = shp, scale = scle)]
 
   if(.return_tibble){
     df <- dplyr::as_tibble(df)

R/random-tidy-general-beta.R

@@ -37,8 +37,8 @@
 #' @param .scale Must be strictly positive.
 #' @param .rate An alternative way to specify the `.scale` parameter.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_generalized_beta()

R/random-tidy-general-pareto.R

@@ -35,8 +35,8 @@
 #' @param .scale Must be positive.
 #' @param .rate An alternative way to specify the `.scale` argument
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_generalized_pareto()

R/random-tidy-geom.R

@@ -31,8 +31,8 @@
 #' @param .n The number of randomly generated points you want.
 #' @param .prob A probability of success in each trial 0 < prob <= 1.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_geometric()

R/random-tidy-hypergeometric.R

@@ -33,8 +33,8 @@
 #' @param .nn The number of black balls in the urn
 #' @param .k The number of balls drawn fro the urn.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_hypergeometric()

R/random-tidy-logistic.R

@@ -32,8 +32,8 @@
 #' @param .location The location parameter
 #' @param .scale The scale parameter
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_logistic()

R/random-tidy-lognormal.R

@@ -32,8 +32,8 @@
 #' @param .meanlog Mean of the distribution on the log scale with default 0
 #' @param .sdlog Standard deviation of the distribution on the log scale with default 1
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_lognormal()

R/random-tidy-negative-binomial.R

@@ -35,8 +35,8 @@
 #' positive, need not be integer.
 #' @param .prob Probability of success on each trial where 0 < .prob <= 1.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_negative_binomial()

R/random-tidy-normal-inverse.R

@@ -31,8 +31,8 @@
 #' @param .shape Must be strictly positive.
 #' @param .dispersion An alternative way to specify the `.shape`.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_inverse_normal()

R/random-tidy-normal.R

@@ -31,8 +31,8 @@
 #' @param .mean The mean of the randomly generated data.
 #' @param .sd The standard deviation of the randomly generated data.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_normal()

R/random-tidy-paralogistic.R

@@ -33,8 +33,8 @@
 #' @param .scale Must be strictly positive.
 #' @param .rate An alternative way to specify the `.scale`
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_paralogistic()

R/random-tidy-pareto-inverse.R

@@ -33,8 +33,8 @@
 #' @param .shape Must be positive.
 #' @param .scale Must be positive.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_inverse_pareto()

R/random-tidy-pareto-single-param.R

@@ -32,8 +32,8 @@
 #' @param .shape Must be positive.
 #' @param .min The lower bound of the support of the distribution.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_pareto1()

R/random-tidy-pareto.R

@@ -32,8 +32,8 @@
 #' @param .shape Must be positive.
 #' @param .scale Must be positive.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_pareto()

R/random-tidy-poisson.R

@@ -32,8 +32,8 @@
 #' @param .n The number of randomly generated points you want.
 #' @param .lambda A vector of non-negative means.
 #' @param .num_sims The number of randomly generated simulations you want.
-#' @param .return_tibble Should the function return a tibble. FALSE will return
-#' a data.table object.
+#' @param .return_tibble A logical value indicating whether to return the result
+#' as a tibble. Default is TRUE.
 #'
 #' @examples
 #' tidy_poisson()