glue-1.2.0.Rmd

---
title: glue 1.2.0
date: '2017-10-31'
slug: glue-1.2.0
author: Jim Hester
categories: [package]
description: >
  glue 1.2.0 is now available on CRAN. glue is designed to make it easy to
  interpolate ("glue") your data into strings.
photo:
  url: http://www.littlerock.af.mil/News/Photos/igphoto/2001697811/
  author: Airman 1st Class Grace Nichols
---

```{r setup, include = FALSE}
knitr::opts_chunk$set(
  comment = "#>",
  collapse = TRUE
)
library(glue)
```

[glue 1.2.0](http://glue.tidyverse.org) is now available on CRAN!
[glue](http://glue.tidyverse.org) is designed to make it easy to interpolate
("glue") your data into strings. Compared to equivalents like `paste()` and
`sprintf()` it is easier to write and less time consuming to maintain. It also
has no non-base dependencies so is easy to include in packages.

Install the latest version with:

```{r, eval = FALSE}
install.packages("glue")
```

glue has three primary functions, `glue()`, `glue_data()` and `collapse()`.
`glue()` works in a similar way to double quotes `"` in a shell or python's
[String Interpolation](https://www.python.org/dev/peps/pep-0498/). You surround
the code you want evaluated by `{}` and the value of the expression is inserted
into the string.

```{r}
name <- "Fred"
age <- 50
anniversary <- as.Date("1991-10-12")
glue('
  My name is {name}
  my age next year is {age + 1}
  my anniversary is {format(anniversary, "%A, %B %d, %Y")}
')
```

glue is also vectorized over its inputs.

```{r}
glue('
  {month.abb} is short for {month.name}
')
```

`glue_data()` works like `glue()`, but instead of looking up its variables
from the calling environment it looks them up from the first argument (usually
a data frame or tibble). This makes `glue_data()` very useful within pipe chains.

```{r}
library(magrittr)
mtcars$model <- rownames(mtcars)
mtcars %>%
  head %>%
  glue_data("The {model} has {gear} gears, {cyl} cylinders, and {hp} horsepower.")
```

`collapse()` is used to combine multiple values into one. The `last` argument
is used to change the separator for the last value.

```{r}
collapse(1:5, ", ", last = ", and ")
```

## glue transformers

New to glue 1.2.0 are transformer functions, which allow you to define custom
behavior for glue functions. For example a `collapse_transformer()` which
automatically collapses any blocks which end with `*`.

```{r}
collapse_transformer <- function(regex = "[*]$", ...) {
  function(code, envir) {
    if (grepl(regex, code)) {
        code <- sub(regex, "", code)
    }
    res <- eval(parse(text = code), envir = envir)
    collapse(res, ...)
  }
}

glue("
  {1:5*}
  {letters[1:5]*}",
  .transformer = collapse_transformer(sep = ", ", last = ", and "))
```

Or an sprintf transformer which lets you use sprintf style numeric formatting with glue.

```{r}
sprintf_transformer <- function(code, envir) {
  m <- regexpr("%.+$", code)
  if (m != -1) {
    format <- regmatches(code, m)
    regmatches(code, m) <- ""
    res <- eval(parse(text = code), envir = envir)
    do.call(sprintf, list(format, res))
  } else {
    eval(parse(text = code), envir = envir)
  }
}

glue_fmt <- function(..., .envir = parent.frame()) {
  glue(..., .transformer = sprintf_transformer, .envir = .envir)
}
glue_fmt("π = {pi%.5f}")
```

## glue_sql()

Also new to glue 1.2.0 is `glue_sql()` and `glue_data_sql()`, which are helper
functions defined with glue transformers to make it easy and safe to construct
SQL statements.

Using `glue_sql()` values are automatically quoted appropriately and variables
can be quoted with backticks.

```{r}
con <- DBI::dbConnect(RSQLite::SQLite(), ":memory:")
colnames(iris) <- gsub("[.]", "_", tolower(colnames(iris)))
DBI::dbWriteTable(con, "iris", iris)
var <- "sepal_width"
tbl <- "iris"
num <- 2
val <- "setosa"
glue_sql("
  SELECT {`var`}
  FROM {`tbl`}
  WHERE {`tbl`}.sepal_length > {num}
    AND {`tbl`}.species = {val}
  ", .con = con)
```

## Other changes

There are many other bug fixes and other minor improvements. You can see a
complete list in the [release
notes](https://github.com/tidyverse/glue/releases/tag/v1.2.0).

A big thanks goes to all the community members who contributed code and opened
issues since the last release!
([\@artemklevtsov](https://github.com/artemklevtsov),
[\@daroczig](https://github.com/daroczig),
[\@DarwinAwardWinner](https://github.com/DarwinAwardWinner),
[\@edarague](https://github.com/edarague),
[\@hadley](https://github.com/hadley),
[\@hughjonesd](https://github.com/hughjonesd),
[\@jennybc](https://github.com/jennybc),
[\@jimhester](https://github.com/jimhester),
[\@jjchern](https://github.com/jjchern), [\@klmr](https://github.com/klmr),
[\@krlmlr](https://github.com/krlmlr), [\@lionel-](https://github.com/lionel-),
[\@mgirlich](https://github.com/mgirlich),
[\@mmuurr](https://github.com/mmuurr), [\@npjc](https://github.com/npjc),
[\@pssguy](https://github.com/pssguy), and
[\@robinsones](https://github.com/robinsones))