Skip to content

Write code faster

Beginners lose time on column names, on the arguments of a function, and on the layout of other people’s code. DataStep Studio for R does these jobs, so you can think about the logic. The keys are on the cheat sheet.

When the session runs, the editor asks it what fits at the cursor: the functions and their arguments, the packages and their functions, the data frames, and the columns of the data frame that a call or a pipe works on, each with its label. Suggestions open as you type, or with Ctrl+Space.

Run programs/22_3_pct.R once. In line 12, put the cursor inside count( and press Ctrl+Space. The first suggestions are the columns of adsl. dsr complete shows the same list in a terminal (here the first 12 of 62):

$ dsr complete programs/22_3_pct.R:12:9 --n 12
programs/22_3_pct.R:12:9: argument in count(), data: adsl (from the session)
STUDYID column "Study Identifier" · character · adsl
USUBJID column "Unique Subject Identifier" · character · adsl
SUBJID column "Subject Identifier for the Study" · character · adsl
SITEID column "Study Site Identifier" · character · adsl
COUNTRY column "Country" · character · adsl
RFSTDTC column "Subject Reference Start Date/Time" · character · adsl
RFENDTC column "Subject Reference End Date/Time" · character · adsl
RFXSTDTC column "Date/Time of First Study Treatment" · character · adsl
RFXENDTC column "Date/Time of Last Study Treatment" · character · adsl
RFPENDTC column "Date/Time of End of Participation" · character · adsl
SCRFDT column "Screen Failure Date" · Date · adsl
FRVDT column "Final Retrieval Visit Date" · Date · adsl
... 50 more (--n)
Signature: count(x, ..., wt = NULL, sort = FALSE, name = NULL) (now: x: A data frame, data frame extension (e.g. a tibble), or a lazy data frame (e.g. from dbplyr or dtplyr).)
COMPLETE: 62 candidates (session)
  • Hover. Rest the mouse on a function: its help. On a column: its label and its missing values.
  • Signature help. Type if_else( and a hint shows each argument with its help.
  • Without a session, completion still knows the project’s functions and the packages’ functions.

Put the cursor on a function and press Shift+F1 (DataStep R: Help for the Function at the Cursor). A side panel shows the help page of the installed version of the package: usage, arguments, value, examples. Run example sends the examples to the session. The Search box finds help pages and vignettes of every installed package; the long guides of admiral are under Vignettes.

F1 stays the command palette. dsr syntax gives the arguments of a function in a terminal, before you write the call:

$ dsr syntax dplyr::count
dplyr::count (version 1.2.1)
count(x, ..., wt = NULL, sort = FALSE, name = NULL)
Count the observations in each group

(The first lines; the rest of the help page follows.)

22.3 Snippets: a clinical program from a short prefix

Section titled “22.3 Snippets: a clinical program from a short prefix”

A snippet is a piece of code that the editor writes for you. Type its prefix and press Tab. The snippets of DataStep Studio for R are tested on the pharmaverse data:

Prefix What it writes
dsr-header a program header: purpose, inputs, outputs, specification, QC
adsl-trt, adsl-saffl the treatment dates and the safety flag from EX (admiral)
adsl-agegr age groups, with their numbers
adae-dates, bds-chg, adlb-param, adtte-param ADAE dates and TRTEMFL, baseline and change, ADLB, a time-to-event parameter (admiral)
t-incidence, t-summary, t-shell, t-tte table programs: subjects with an event, summary statistics, a table from its shell, Kaplan-Meier
pry-when a stop for one subject (chapter 24)
test-that a testthat test (chapter 26)

This program works, but nobody wants to read it:

programs/22_1_messy.R
# Chapter 22: a program that works, in a layout nobody wants to read
library(dplyr)
adsl<-pharmaverseadam::adsl|>filter(SAFFL=="Y")
adsl|>group_by(TRT01A)|>summarise(n=n(),mean_age=round(mean(AGE),1),
over_80=sum(AGE>80))
#> # A tibble: 3 x 4
#> TRT01A n mean_age over_80
#> <chr> <int> <dbl> <int>
#> 1 Placebo 86 75.2 30
#> 2 Xanomeline High Dose 72 73.8 12
#> 3 Xanomeline Low Dose 96 76 35

Press Ctrl+Shift+A (Mac: Cmd+Shift+A), Format R Program. With a selection, only those lines change. Shift+Alt+F, VS Code’s own Format Document, uses the same engine. The file then has the tidyverse layout:

# Chapter 22: a program that works, in a layout nobody wants to read
library(dplyr)
adsl <- pharmaverseadam::adsl |> filter(SAFFL == "Y")
adsl |>
group_by(TRT01A) |>
summarise(
n = n(), mean_age = round(mean(AGE), 1),
over_80 = sum(AGE > 80)
)

Before it writes, the formatter compares the program before and after, token by token. A token is one word of the language: a name, an operator, a number, a string. Only when every token and every comment are the same, the file is written. This is the proof that only the layout changed. dsr format --check gives the proof and writes nothing:

$ dsr format programs/22_1_messy.R --check
would change: programs/22_1_messy.R (proof: only the layout changed: the 52 tokens and 1 comments are the same)
FORMAT: WOULD CHANGE (1 files: 1 would be formatted, 0 already in the style, 0 refused; scope line_breaks: the layout only; indent 2 (styler's defaults), 1 would change; line limit 80: 0 breaks by the wrap step, 0 lines over it before, 0 would stay)

dsr format programs/22_1_messy.R writes the file. Cmd/Ctrl+Z takes a format back in the editor. The setting dsr.format.scope says how much may change; the default line_breaks changes the layout only.

Style is how code looks: spaces, names, line length. R’s usual standard is the tidyverse style guide. dsr style (or DataStep R: Check the Style of This File) lists where a file leaves it, each finding with its line and its fix. In the editor, the same findings are in the Problems panel as information.

$ dsr style programs/22_1_messy.R
programs/22_1_messy.R:3: spacing: no space around <- | fix: dsr format (in the editor: Format Document) puts the spaces; it proves that only the layout changes
programs/22_1_messy.R:4: spacing: no space around |> | fix: as for programs/22_1_messy.R:3
programs/22_1_messy.R:5: spacing: no space around = | fix: as for programs/22_1_messy.R:3
programs/22_1_messy.R:4: lintr:pipe_continuation: Put a space before `|>` and a new line after it, unless the full pipeline fits on one line. | fix: see ?lintr::pipe_continuation_linter
programs/22_1_messy.R:5: lintr:indentation: Hanging indent should be 34 spaces but is 2 spaces. | fix: see ?lintr::indentation_linter
STYLE: FINDINGS (5 findings in 1 of 1 file: spacing 3, lintr:indentation 1, lintr:pipe_continuation 1; the tidyverse style guide, line length 80; with lintr's default linters; a style finding is not an error. dsr format fixes the layout, dsr style --fix the safe ones)

A style finding is never an error. Format fixes the layout. A function without a roxygen header (the #' comment lines above a function, chapter 19) gets the quick fix Insert roxygen header: put the cursor on the function and press Ctrl+. (Mac: Cmd+.).

Lint reads the code without running it and finds real problems: a function that does not exist, an argument a function does not have, a variable that nobody made. This program has two typos:

programs/22_2_typo.R
# Chapter 22: two typos that dsr lint finds before you run the program
library(dplyr)
adsl_saf <- pharmaverseadam::adsl |>
filter(SAFFL == "Y")
adsl_saf |>
group_by(TRT01A) |>
summarise(subjects = n_distnct(USUBJID))
#> Error in `summarise()`:
#> i In argument: `subjects = n_distnct(USUBJID)`.
#> i In group 1: `TRT01A = "Placebo"`.
#> Caused by error in `n_distnct()`:
#> ! could not find function "n_distnct"

The run stopped at the first typo. The file has one more line, nrow(adsl_safe), which R never reached. dsr lint (or DataStep R: Lint the Open R Files) finds both typos at once, with a suggestion for each:

$ dsr lint programs/22_2_typo.R
programs/22_2_typo.R:10: function: no visible global function definition for ‘n_distnct’ (did you mean n_distinct?)
programs/22_2_typo.R:12: variable: no visible binding for global variable ‘adsl_safe’ (did you mean adsl_saf?)
LINT: PROBLEMS (1 files, with lintr)

On a long program, a typo near the end costs you a whole run. Lint costs a second. Lint each program before you run it, and always before you hand it to QC.

A study has many files: functions in one file, a program per data set, a program per table. These keys move between them. They read the files; they need no run.

Key (Mac: Cmd for Ctrl) What it does
Ctrl+click, Ctrl+Alt+G go to the definition of a function, a data frame or a column
Alt+F12 peek: the definition in place, under the line
Shift+F12 every use, in every R file of the project
Ctrl+Alt+U Find Every Use: the definition, the uses and, apart, the possible uses
Shift+F2 rename a function or a variable, with a proof
Ctrl+Alt+O the functions, sections and data frames of this file
Ctrl+T a function or a data frame of the whole project, by part of its name
Ctrl+Alt+J back to the line you came from

In VS Code, F12 is Go to Definition and F2 is Rename. In R files, DataStep Studio for R gives F12 to the data viewer and F2 to clearing the log, as in the SAS tool. So Rename is on Shift+F2 and Go to Definition on Ctrl+click.

Try them on this program:

programs/22_3_pct.R
# Chapter 22: a small function, used twice: go to its definition, find its uses, rename it
library(dplyr)
pct_subjects <- function(n, total) {
round(100 * n / total, 1)
}
adsl <- pharmaverseadam::adsl |>
filter(SAFFL == "Y")
adsl |>
count(TRT01A, SEX) |>
mutate(pct = pct_subjects(n, sum(n)), .by = TRT01A)
#> # A tibble: 6 x 4
#> TRT01A SEX n pct
#> <chr> <chr> <int> <dbl>
#> 1 Placebo F 53 61.6
#> 2 Placebo M 33 38.4
#> 3 Xanomeline High Dose F 35 48.6
#> 4 Xanomeline High Dose M 37 51.4
#> 5 Xanomeline Low Dose F 55 57.3
#> 6 Xanomeline Low Dose M 41 42.7
adsl |>
count(AGEGR1) |>
mutate(pct = pct_subjects(n, sum(n)))
#> # A tibble: 2 x 3
#> AGEGR1 n pct
#> <chr> <int> <dbl>
#> 1 18-64 33 13
#> 2 >64 221 87

Each move has a dsr command. Where is pct_subjects defined, and where is it used?

$ dsr def pct_subjects
programs/22_3_pct.R:4 function pct_subjects(n, total)
4 pct_subjects <- function(n, total) {
5 round(100 * n / total, 1)
6 }
DEF: FOUND | pct_subjects is a function: programs/22_3_pct.R:4
$ dsr refs pct_subjects
pct_subjects: function, the project
defined programs/22_3_pct.R:4:1 pct_subjects <- function(n, total) {
use programs/22_3_pct.R:13:16 mutate(pct = pct_subjects(n, sum(n)), .by = TRT01A)
use programs/22_3_pct.R:17:16 mutate(pct = pct_subjects(n, sum(n)))
REFS: 2 uses in 1 file, 1 definition, 0 possible (pct_subjects: function)

A rename changes the name in every file where it means that function, and nothing else. --check shows the plan and writes nothing:

$ dsr rename pct_subjects pct_of_n --check
would change programs/22_3_pct.R:4:1 pct_of_n <- function(n, total) {
would change programs/22_3_pct.R:13:16 mutate(pct = pct_of_n(n, sum(n)), .by = TRT01A)
would change programs/22_3_pct.R:17:16 mutate(pct = pct_of_n(n, sum(n)))
RENAME: WOULD CHANGE | pct_subjects -> pct_of_n (function, the project): 3 tokens in 1 file (proof: 3 tokens changed (pct_subjects to pct_of_n) in 1 file; the other 82 tokens, every comment and every string are the same; R's parser confirms it)

A rename of a column is refused: which data frame a column name belongs to is decided only when the code runs. Use Find Every Use to see the possible places, and change them yourself.

Two keys show or hide the arguments of a call. They are in the Professional edition, free during Early Access until 17 February 2027.

  • Ctrl+Shift+E, Enlarge Function Call: every argument on its own line, with its name, and the arguments you did not give with their defaults.
  • Ctrl+Shift+O, Optimize Function Call: the arguments that have their default value out, then the call on one line when it fits.

Both are written only with a proof that the call means the same. In line 14 of 20_1_weight.R, enlarge the if_else() call:

$ dsr call enlarge programs/20_1_weight.R:14:20 --check
mutate(WGTGR = if_else(
condition = WEIGHTBL >= 70,
true = ">=70 kg",
false = "<70 kg",
missing = NULL,
ptype = NULL
))
CALL: WOULD BE ENLARGED (programs/20_1_weight.R:14: if_else() of the package dplyr: 3 names written, 2 defaults added (missing, ptype); `...` is not added; size: not added (the default is computed inside the function, or the function asks missing()); proof: the matched call with the defaults filled in is the same before and after; the comments and the code around the call are the same; line limit 80)

Now you see an argument you did not know: missing. It is the value for a row where the condition is NA. In chapter 20, the subject without a baseline weight got NA because of it. missing = "no weight" would give that subject a group of its own.

  • Ctrl+Space completes the columns of the data frame you work on, with their labels. Shift+F1 opens the help of the installed version.
  • Snippets (adsl-agegr, t-incidence, test-that, …) write tested clinical code from a prefix and Tab.
  • Ctrl+Shift+A formats a program to the tidyverse layout, and writes only with the proof that only the layout changed.
  • dsr style lists where the code leaves the style guide. dsr lint finds typos and missing functions before you run.
  • Ctrl+click, Shift+F12 and Shift+F2 move to a definition, list the uses and rename with a proof.
  • Ctrl+Shift+E and Ctrl+Shift+O enlarge and optimize a call (Professional).

Exercise 22-1. Format 22_1_messy.R with dsr format --check. How many tokens and comments does the proof compare?

Exercise 22-2. Fix the two typos of 22_2_typo.R. What does dsr lint say now, and how many subjects are in each arm?

Exercise 22-3. Rename pct_subjects to pct_of_n with --check. How many tokens would change, and how many stay the same?

Exercise 22-4. Optimize the if_else() call in line 14 of 20_1_weight.R with dsr call optimize ... --check. Does anything change? Why?

The answers are in Answers, Part VI.

Found an error or have a question? Write to datastep.studio@gmail.com· Every program runs in DataStep Studio for R, the R extension for VS Code and Positron.

Data: the CDISC pilot study (cdiscpilot01), from the R packagespharmaversesdtm andpharmaverseadam (Apache License 2.0).

R is a project of the R Foundation. SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. This book is not affiliated with or endorsed by SAS Institute Inc. or the R Foundation. CDISC, SDTM and ADaM are trademarks of the Clinical Data Interchange Standards Consortium.

© 2026 DataStep Labs. All rights reserved.