In ashiklom/metar: Add Metadata to your Data

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

metar

metar combines several utilities, including many packages of the core tidyverse, the sticky package for persistent attributes, and the data.table package for fast I/O, to provide a comprehensive interface for working with data and its metadata. Currently the package provides utilities for:

• Reading from CSVY files, which provide a convenient way of storing tabular data and its metadata together while still being readable by common CSV-reading utilities.
• Adding metadata to in-memory data.frames with a dplyr-like syntax

Installation

This package is not on CRAN (yet!), but is easy enough to install from GitHub:

# install.packages("devtools")
devtools::install_github("ashiklom/metar")

Usage

Add metadata directly to objects:

library(metar)

obj <- rnorm(100)
metadata(obj) <- list(title = "Normal draws", description = "Some random normal draws")
metadata(obj)
metadata(obj, "title", "description")

You can also use the pipe-friendly add_metadata:

obj2 <- runif(100) %>%
  add_metadata(title = "Normal draws") %>%
  # Metadata can be added or overwritten
  add_metadata(author = "RNG", title = "Just kidding, they're uniform draws") %>%
  # Metadata can be any R object
  add_metadata(range = list(min = 0, max = 1))
metadata(obj2)
metadata(obj2, "range")

Add metadata to each column of the iris dataset:

library(dplyr)

iris_md <- iris %>%
  as_tibble() %>%
  add_column_metadata(
    # Metadata for individual columns
    Sepal.Length = list(title = "Sepal length", unit = "cm"),
    Sepal.Width = list(title = "Sepal width", unit = "in"),
    # .root sets metadata for the data frame as a whole
    .root = list(description = "The famous Iris dataset", author = "Not Alexey Shiklomanov")
  )

Metadata are stored as attributes.

attr(iris_md, "description")
attr(iris_md, "author")
attr(iris_md$Sepal.Length, "title")
attributes(iris_md$Sepal.Width)

These attributes are made persistent through many operations by the sticky package.

metadata(iris_md)
metadata(iris_md[1, ], "description")
metadata(iris_md$Sepal.Length[1:5], "title")

The CSVY format is just a CSV file with a YAML header.

system.file("examples/example1.csvy", package = "metar") %>%
  readLines() %>%
  writeLines()

If read in with the read_csvy function, all metadata will be automatically applied:

dat <- read_csvy(system.file("examples/example1.csvy", package = "metar"))
dat
metadata(dat)
lapply(dat, metadata)

Note that because var2 was specified as an integer class, it was also read in as an integer rather than a numeric.

class(dat$var2)

ashiklom/metar documentation built on May 25, 2019, 2:26 p.m.