Gabriel Becker and Martin Maechler
The useR! 2021 conference Code of Conduct applies to all aspects of this tutorial and will be enforced throughout. By joining the tutorial session you agree to abide by the code of conduct at all times in both the main tutorial call and all break-out 'rooms'.
Please bring any violations to the code of conduct to the immediate attention of Gabriel, Martin, or one of the helpers in attendence.
The tutorial - like all of the live aspects of the useR! 2021 virtual conference - will take place on Zoom. Unlike most sessions in the conference, the tutorial will be a zoom meeting, not a webinar. As such it cannot be attended via web browser. Please ensure you have an up-to-date Zoom client installed on your device prior to the start of the tutorial.
Portions of this tutorial will be participatory so a working zoom-compatible microphone and a location where you can speak loud enough to be understood are required.
A working camera and internet connection stable enough to transmit video are strongly preferred, if possible.
Part or all of this tutorial will be recorded and made available to the wider R community. By joining and remaining in the session you agree to have this tutorial and your participation in it recorded and the resulting video made publicly available.
This tutorial will assume certain baseline knowledge about how R works in order to effectively use the short amount of time we have.
Please read or otherwise be familiar with the subjects covered in the following:
Accessing R Sources - Uwe Ligges, an excerpt from R News 2006-4; R Help Desk, p.43--45.
Writing R Documentation (Rd Syntax for package authors)
Parsing Rd Files - Duncan Murdoch (Rd Syntax Specification section)
Help pages UseMethod and methods
(S3) Object oriented Programming
Search Paths (and the following Namespaces section)
For for the practical portions of this tutorial, participants will use use Eddelbuettel and Boettiger's rocker project to step into older versions of R to explore real bugs which have since been fixed.
Prior to the start of the session please have docker installed and a rocker image version with R 3.3.2 pulled and tested (rocker/r-ver:3.3.2 or rocker/rstudio:3.3.2).
Participants unable to run docker on the device they will be joining the session from will be placed within breakout groups where at a least one participant does have the setup working, and can share their screen, but working along locally is still preferable.
Note you can also use podman instead of docker, notably on
Fedora/Redhat/... Linux systems where it is well supported.
In the terminal, simply replace the word docker by podman.
If using the rstudio image, please read and understand the instructions for using the container ( https://hub.docker.com/r/rocker/rstudio) and come ready and able to use the container.
If for some reason you choose not to use the rstudio based images, ensure your container has some other progam usable as an IDE (e.g., emacs with ess) installed.
Configure a volume for your container, like so (see full documentation for volumes here: https://docs.docker.com/storage/volumes/)
docker volume create r-source
Ensure your volume was created using docker volume ls
Gabriels-MacBook-Pro:rtables_paper gabrielbecker$ docker volume ls
DRIVER VOLUME NAME
<snip>
local r-source
Then start a shell in your docker container with the volume mounted
docker run -it --mount src=r-source,target=/r-source rocker/rstudio:3.3.2 bash
root@788c85efe291:/#
By doing ls we can see that our mounted volume is there:
root@788c85efe291:/# ls
bin boot dev etc home init lib lib64 media mnt opt proc root r-source run sbin srv sys tmp usr var
Then navigate to that directory and use wget and tar to download and untar the R 3.3.2 source tarball ( https://cloud.r-project.org/src/base/R-3/R-3.3.2.tar.gz ) onto it.
root@93dde2224f94:/r-source# wget https://cloud.r-project.org/src/base/R-3/R-3.3.2.tar.gz
--2021-07-02 17:20:01-- https://cloud.r-project.org/src/base/R-3/R-3.3.2.tar.gz
Resolving cloud.r-project.org (cloud.r-project.org)... 204.246.191.121, 204.246.191.87, 204.246.191.77, ...
Connecting to cloud.r-project.org (cloud.r-project.org)|204.246.191.121|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 29440670 (28M) [application/x-gzip]
Saving to: ‘R-3.3.2.tar.gz’
R-3.3.2.tar.gz 100%[================================================================================================================================================>] 28.08M 9.03MB/s in 3.1s
2021-07-02 17:20:04 (9.03 MB/s) - ‘R-3.3.2.tar.gz’ saved [29440670/29440670]
root@93dde2224f94:/r-source# tar -xvf R-3.3.2.tar.gz
R-3.3.2/
R-3.3.2/ChangeLog
R-3.3.2/config.site
R-3.3.2/configure
<snip>
We will explore and use portions of these sources in the practical portions of the tutorial.
Ensure these files are persistent by closing the docker container (e.g. via exit, and invoking it again to ensure the files are still there:
root@788c85efe291:/# exit
exit
Gabriels-MacBook-Pro:rtables_paper gabrielbecker$ cd ~
Gabriels-MacBook-Pro:~ gabrielbecker$ docker run -it --mount src=r-source,target=/r-source rocker/rstudio:3.3.2 bash
root@15a8d340686a:/# cd r-source/
root@15a8d340686a:/r-source# ls
R-3.3.2 R-3.3.2.tar.gz
When invoking the Rstudio container as described at https://hub.docker.com/r/rocker/rstudio) be sure to remember to add --mount src=r-source,target=/r-source to the invocation.
joe <- person("Joe", "Schmo", email = c("[email protected]", "[email protected]"))
str(joe$email)
## chr [1:2] "[email protected]" "[email protected]"
joe_text <- format(joe)
print(joe_text)
## [1] "Joe Schmo <[email protected], [email protected]>"
joe_new <- as.person(joe_text)
str(joe_new$email)
## chr "[email protected], [email protected]"
library(survival)
library(relsurv)
data("slopop")
is.ratetable(slopop)
# [1] TRUE
is.ratetable(slopop, verbose = TRUE)
# [1] "wrong length for cutpoints 3"
d <- as.POSIXct("2016-06-08 14:21", tz="US/Pacific") + as.difftime(2^(-2:8), units="mins")
str(d)
# POSIXct[1:11], format: "2016-06-08 14:21:15" "2016-06-08 14:21:30" ...
str(diff(d))
#Class 'difftime' atomic [1:10] 15 30 60 120 240 480 960 1920 3840 7680
# ..- attr(*, "units")= chr "secs"
str(diff(diff(d)))
#Class 'difftime' num [1:9] 15 30 60 120 240 480 960 1920 3840
addmargins() fails if supplied functions are not defined in the stats namespace or parent environment thereof
local({
mB <- structure(c(16, 26, 27, 20, 24, 20, 19, 25, 40, 46, 46, 45),
.Dim = c(4L, 3L),
.Dimnames = list(Sea = c("Black", "Dead", "Red", "White"),
Bee = c("Buzz", "Hum", "Total")),
class = c("table", "matrix"))
sqsm <- function(x) sum(x)^2/100
addmargins(mB, 1, list(list(All = sum, N = sqsm)))
})
data(iris)
lapply(iris, class)
Species2 <- iris$Species
Sepal.Length2 <- iris$Sepal.Length
class(Species2) <- c("some_class", class(Species2))
class(Sepal.Length2) <- c("some_class", class(Sepal.Length2))
attr(Species2, "some_attr") <- "some_attr_val"
attr(Sepal.Length2, "some_attr") <- "some_attr_val"
iris$Species2 <- Species2
iris$Sepal.Length2 <- Sepal.Length2
lapply(iris, class)
lapply(iris, attributes)
iris2 <- iris[c(2:5), ] # row-wise subsetting
lapply(iris2, class) # Species2: class c("some_class", "factor"), Sepal.Length2 class stripped to numeric
lapply(iris2, attributes) # all (incl. Species2): "some_attr" is stripped
library("mgcv") # or "survival", or just loadNamespace("Matrix")
f <- factor(1:10)
debugcall(summary(f))
- Search for "Should last default to" / "for substring()" in the R-devel list archives
- Look at the first 2--3 mails; what do you think? how would you solve it?