In jmbarbone/jordan: Miscellaneous, Analytic R Kernels

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

set.seed(42)

mark

Miscellaneous, Analytic R Kernels

An R package with a set of general use functions for data analytics. This is developed mostly for personal use and has no real goal other than to limit the time I spend searching where I did that thing that I think I could use again because it worked well but this problem might be slightly different and I know I had to change it before.

Some parts happily ripped from and (hopefully) credited to others.

Installation

You can download the current CRAN version with:

install.packages("mark")

You can the development version from GitHub with:

remotes::install_github("jmbarbone/mark")

Select examples

This package contains a many variety of functions, some useful, some not so much. Below are a selection of a few functions that could potential be useful for others:

library(mark)

Get dates from sloppy entries:

bad_dates <- c("2020 Dec 8th", "1970 May", "??", "1984 UNK UN")
date_from_partial(bad_dates)
date_from_partial(bad_dates, method = "max")
date_from_partial(c("May 2000", "08Dec2020"), format = "dmy")

Slice strings:

x <- stringi::stri_rand_lipsum(1)
str_slice(x, n = 50L)
str_slice_by_word(x)

Read in bibliographies:

file <- system.file("extdata", "example-bib.txt", package = "mark")
bib <- read_bib(file)
tibble::as_tibble(bib)

More matching:

1:10 %out% c(1, 3, 5, 9) # opposite of %in% 
letters[1:5] %wo% letters[3:7]
letters[1:5] %wi% letters[3:7]

Small functions for working with data.frames:

x <- list(a = 1:5, b = letters[1:5])
quick_df(x)

vector2df(x[["b"]], name = NULL)

quick_dfl(a = 1:3, b = list(1:5, 6:10, 11:15))

Counts and proportions:

set.seed(42)
x <- sample(1:5, 20, TRUE, 5:1/2)
counts(x)
props(x)

df <- as.data.frame(matrix(sample(1:2, 60, TRUE), byrow = TRUE, ncol = 3))
counts(df, c("V1", "V2"))
props(df, 1:3)

Date time differences:

x <- as.POSIXlt("2021-02-13 05:02:30", tz = "America/New_York") + c(0, -1, 2) * 3600 * 24
y <- as.POSIXlt("2020-02-13 05:02:30", tz = "America/New_York") + c(0, -2, 4) * 3600 * 24

# comparison with base::difftime() (note the order of x and y)
difftime(y, x, units = "days")
diff_time_days(x, y)

difftime(y, x, units = "secs")
diff_time_secs(x, y)

# Year (by days, months, etc)
diff_time_years(x, y)
diff_time_myears(x, y)

# Set time zones
diff_time_hours(x, y, "GMT", "America/New_York")                         
diff_time_hours(x, x, "GMT", c("America/Los_Angeles", "America/New_York", "Europe/London")) # note x, x
diff_time_days(x, y, NULL, 31536000)

Simple factors:

fact(c("a", "c", NA, "a", "b", NA, "a", "c")) # no sorting
fact(c(-1, 5, 2, NA, 3))                      # sorting
fact(c(NA, FALSE, TRUE, FALSE, TRUE, NA))     # fixed