1 <?xml version="1.0"?> 2 <html xmlns="http://www.w3.org/1999/xhtml"> 3 <head> 4 <meta http-equiv="Content-Type" content="text/html; charset=UTF-8"/> 5 <link rel="stylesheet" type="text/css" href="styleguide.css"/> 6 <title>Google's R Style Guide</title> 7 </head> 8 9 <body> 10 11 <h1>Google's R Style Guide</h1> 12 13 <p> 14 R is a high-level programming language used primarily for 15 statistical computing and graphics. The goal of the R 16 Programming Style Guide is to make our R code easier to read, 17 share, and verify. The rules below were designed in 18 collaboration with the entire R user community at Google. 19 20 </p> 21 22 23 24 25 26 27 <h2>Summary: R Style Rules</h2> 28 29 <ol> 30 <li><a href="#filenames">File Names</a>: end in <code>.R</code></li> 31 <li><a href="#identifiers">Identifiers</a>: <code>variable.name</code> 32 (or <code>variableName</code>), 33 <code>FunctionName</code>, <code>kConstantName</code></li> 34 <li><a href="#linelength">Line Length</a>: maximum 80 characters</li> 35 <li><a href="#indentation">Indentation</a>: two spaces, no tabs</li> 36 <li><a href="#spacing">Spacing</a></li> 37 <li><a href="#curlybraces">Curly Braces</a>: first on same line, last on 38 own line</li> 39 <li><a href="#else">else</a>: Surround else with braces </li> 40 <li><a href="#assignment">Assignment</a>: use <code><-</code>, not 41 <code>=</code></li> 42 <li><a href="#semicolons">Semicolons</a>: don't use them</li> 43 <li><a href="#generallayout"> General Layout and Ordering</a></li> 44 <li><a href="#comments"> Commenting Guidelines</a>: all comments begin 45 with <code>#</code> followed by a space; inline comments need two 46 spaces before the <code>#</code></li> 47 <li><a href="#functiondefinition">Function Definitions and Calls</a></li> 48 <li><a href="#functiondocumentation"> Function Documentation</a></li> 49 <li><a href="#examplefunction"> Example Function</a></li> 50 <li><a href="#todo"> TODO Style</a>: <code>TODO(username)</code></li> 51 </ol> 52 53 <h2>Summary: R Language Rules</h2> 54 <ol> 55 <li><a href="#attach"> <code>attach</code></a>: avoid using it</li> 56 <li><a href="#functionlanguage"> Functions</a>: 57 errors should be raised using <code>stop()</code></li> 58 <li><a href="#object"> Objects and Methods</a>: avoid S4 objects and 59 methods when possible; never mix S3 and S4 </li> 60 </ol> 61 62 <h3>Notation and Naming</h3> 63 64 <h4 id="filenames">File Names</h4> 65 <p> 66 File names should end in <code>.R</code> and, of course, be 67 meaningful. 68 <br/> GOOD: <code>predict_ad_revenue.R</code> 69 <br/> BAD: <code><span style="color:red">foo.R</span></code> 70 </p> 71 72 <h4 id="identifiers">Identifiers</h4> 73 <p> 74 Don't use underscores ( <code>_</code> ) or hyphens 75 ( <code>-</code> ) in identifiers. 76 Identifiers should be named according to the following conventions. 77 The preferred form for variable names is all lower case 78 letters and words separated with dots 79 (<code>variable.name</code>), but <code>variableName</code> 80 is also accepted; 81 function names have initial capital letters and no dots 82 (<code>FunctionName</code>); 83 constants are named like functions but with an initial 84 <code>k</code>. 85 </p> 86 <ul> 87 <li><code>variable.name</code> is preferred, <code>variableName</code> is accepted 88 <br/> GOOD: <code>avg.clicks</code> 89 <br/> OK: <code>avgClicks</code> 90 <br/> BAD: <code><span style="color:red">avg_Clicks</span></code> 91 </li> 92 <li><code>FunctionName </code> 93 <br/> GOOD: <code>CalculateAvgClicks</code> 94 <br/> BAD: <code><span style="color:red">calculate_avg_clicks 95 </span></code>, 96 <code><span style="color:red">calculateAvgClicks</span></code> 97 <br/> Make function names verbs. 98 <br/><em>Exception: When creating a classed object, the function 99 name (constructor) and class should match (e.g., lm).</em></li> 100 <li><code>kConstantName </code></li> 101 </ul> 102 103 104 <h3>Syntax</h3> 105 106 <h4 id="linelength">Line Length</h4> 107 <p> 108 The maximum line length is 80 characters. 109 </p> 110 111 <h4 id="indentation">Indentation</h4> 112 <p> 113 When indenting your code, use two spaces. Never use tabs or mix 114 tabs and spaces. 115 <br/><em>Exception: When a line break occurs inside parentheses, 116 align the wrapped line with the first character inside the 117 parenthesis.</em> 118 </p> 119 120 <h4 id="spacing">Spacing</h4> 121 <p> 122 Place spaces around all binary operators (<code>=</code>, 123 <code>+</code>, <code>-</code>, <code><-</code>, etc.). 124 <br/><em> Exception: Spaces around <code>=</code>'s are 125 optional when passing parameters in a function call.</em> 126 </p> 127 <p> 128 Do not place a space before a comma, but always place one after a 129 comma. 130 <br/><br/> GOOD: 131 </p> 132 <pre>tab.prior <- table(df[df$days.from.opt < 0, "campaign.id"]) 133 total <- sum(x[, 1]) 134 total <- sum(x[1, ])</pre> 135 <p> 136 BAD: 137 </p> 138 <pre><span style="color:red">tab.prior <- table(df[df$days.from.opt<0, "campaign.id"]) # Needs spaces around '<' 139 tab.prior <- table(df[df$days.from.opt < 0,"campaign.id"]) # Needs a space after the comma 140 tab.prior<- table(df[df$days.from.opt < 0, "campaign.id"]) # Needs a space before <- 141 tab.prior<-table(df[df$days.from.opt < 0, "campaign.id"]) # Needs spaces around <- 142 total <- sum(x[,1]) # Needs a space after the comma 143 total <- sum(x[ ,1]) # Needs a space after the comma, not before</span> 144 </pre> 145 <p> 146 Place a space before left parenthesis, except in a function call. 147 </p> 148 <p> 149 GOOD: 150 <br/><code>if (debug)</code> 151 </p> 152 <p> 153 BAD: 154 <br/><code><span style="color:red">if(debug)</span></code> 155 </p> 156 <p> 157 Extra spacing (i.e., more than one space in a row) is okay if it 158 improves alignment of equals signs or arrows (<code><-</code>). 159 </p> 160 <pre>plot(x = x.coord, 161 y = data.mat[, MakeColName(metric, ptiles[1], "roiOpt")], 162 ylim = ylim, 163 xlab = "dates", 164 ylab = metric, 165 main = (paste(metric, " for 3 samples ", sep = ""))) 166 </pre> 167 <p> 168 Do not place spaces around code in parentheses or square brackets. 169 <br/><em> Exception: Always place a space after a comma.</em> 170 </p> 171 <p> 172 GOOD:</p><pre>if (debug) 173 x[1, ]</pre> 174 <p> 175 BAD:</p><pre><span style="color:red">if ( debug ) # No spaces around debug 176 x[1,] # Needs a space after the comma </span></pre> 177 178 <h4 id="curlybraces">Curly Braces</h4> 179 <p> 180 An opening curly brace should never go on its own line; a closing 181 curly brace should always go on its own line. You may omit curly 182 braces when a block consists of a single statement; however, you 183 must <em>consistently</em> either use or not use curly braces for 184 single statement blocks. 185 </p> 186 <pre> 187 if (is.null(ylim)) { 188 ylim <- c(0, 0.06) 189 }</pre> 190 <p> 191 xor (but not both) 192 </p> 193 <pre> 194 if (is.null(ylim)) 195 ylim <- c(0, 0.06)</pre> 196 <p> 197 Always begin the body of a block on a new line. 198 </p> 199 <p> 200 BAD: 201 <br/><code><span style="color:red"> if (is.null(ylim)) 202 ylim <- c(0, 0.06)</span></code> 203 <br/><code><span style="color:red"> if (is.null(ylim)) 204 {ylim <- c(0, 0.06)} </span></code> 205 </p> 206 207 <h4 id="else">Surround else with braces</h4> 208 <p> 209 An <code>else</code> statement should always be surrounded on the 210 same line by curly braces.</p> 211 <pre> 212 if (condition) { 213 one or more lines 214 } else { 215 one or more lines 216 } 217 </pre> 218 <p> 219 BAD:<br/> 220 </p> 221 <pre style="color:red"> 222 if (condition) { 223 one or more lines 224 } 225 else { 226 one or more lines 227 } 228 </pre> 229 <p> 230 BAD:<br/> 231 </p> 232 <pre style="color:red"> 233 if (condition) 234 one line 235 else 236 one line 237 </pre> 238 239 <h4 id="assignment">Assignment</h4> 240 <p> 241 Use <code><-</code>, not <code>=</code>, for assignment. 242 </p> 243 <p> 244 GOOD: 245 <br/><code> x <- 5 </code> 246 </p> 247 <p> 248 BAD: 249 <br/><code><span style="color:red"> x = 5</span></code> 250 </p> 251 <h4 id="semicolons">Semicolons</h4> 252 <p> 253 Do not terminate your lines with semicolons or use semicolons to 254 put more than one command on the same line. (Semicolons are not 255 necessary, and are omitted for consistency with other Google style 256 guides.) 257 </p> 258 259 260 <h3> Organization </h3> 261 <h4 id="generallayout">General Layout and Ordering</h4> 262 <p> 263 If everyone uses the same general ordering, we'll be able to 264 read and understand each other's scripts faster and more easily. 265 </p> 266 <ol> 267 <li>Copyright statement comment </li> 268 <li>Author comment</li> 269 <li>File description comment, including purpose of 270 program, inputs, and outputs</li> 271 <li><code>source()</code> and <code>library()</code> statements</li> 272 <li>Function definitions</li> 273 <li>Executed statements, if applicable (e.g., 274 <code> print</code>, <code>plot</code>)</li> 275 </ol> 276 <p> 277 Unit tests should go in a separate file named 278 <code>originalfilename_test.R</code>. 279 </p> 280 <h4 id="comments">Commenting Guidelines</h4> 281 <p> 282 Comment your code. Entire commented lines should begin with 283 <code>#</code> and one space. 284 </p> 285 <p> 286 Short comments can be placed after code preceded by two spaces, 287 <code>#</code>, and then one space. 288 </p> 289 <pre># Create histogram of frequency of campaigns by pct budget spent. 290 hist(df$pct.spent, 291 breaks = "scott", # method for choosing number of buckets 292 main = "Histogram: fraction budget spent by campaignid", 293 xlab = "Fraction of budget spent", 294 ylab = "Frequency (count of campaignids)") 295 </pre> 296 <h4 id="functiondefinition">Function Definitions and 297 Calls</h4> 298 <p> 299 Function definitions should first list arguments without default 300 values, followed by those with default values. 301 </p> 302 <p> 303 In both function definitions and function calls, multiple 304 arguments per line are allowed; line breaks are only allowed 305 between assignments. 306 <br/>GOOD: 307 </p> 308 <pre>PredictCTR <- function(query, property, num.days, 309 show.plot = TRUE) 310 </pre> 311 BAD: 312 <pre><span style="color:red">PredictCTR <- function(query, property, num.days, show.plot = 313 TRUE) 314 </span></pre> 315 <p> Ideally, unit tests should serve as sample function calls (for 316 shared library routines). 317 </p> 318 <h4 id="functiondocumentation">Function Documentation</h4> 319 <p> Functions should contain a comments section immediately below 320 the function definition line. These comments should consist of a 321 one-sentence description of the function; a list of the function's 322 arguments, denoted by <code>Args:</code>, with a description of 323 each (including the data type); and a description of the return 324 value, denoted by <code>Returns:</code>. The comments should be 325 descriptive enough that a caller can use the function without 326 reading any of the function's code. 327 </p> 328 329 <h4 id="examplefunction">Example Function</h4> 330 <pre> 331 CalculateSampleCovariance <- function(x, y, verbose = TRUE) { 332 # Computes the sample covariance between two vectors. 333 # 334 # Args: 335 # x: One of two vectors whose sample covariance is to be calculated. 336 # y: The other vector. x and y must have the same length, greater than one, 337 # with no missing values. 338 # verbose: If TRUE, prints sample covariance; if not, not. Default is TRUE. 339 # 340 # Returns: 341 # The sample covariance between x and y. 342 n <- length(x) 343 # Error handling 344 if (n <= 1 || n != length(y)) { 345 stop("Arguments x and y have different lengths: ", 346 length(x), " and ", length(y), ".") 347 } 348 if (TRUE %in% is.na(x) || TRUE %in% is.na(y)) { 349 stop(" Arguments x and y must not have missing values.") 350 } 351 covariance <- var(x, y) 352 if (verbose) 353 cat("Covariance = ", round(covariance, 4), ".\n", sep = "") 354 return(covariance) 355 } 356 </pre> 357 358 <h4 id="todo">TODO Style</h4> 359 360 <p> 361 Use a consistent style for TODOs throughout your code. 362 <br/><code>TODO(username): Explicit description of action to 363 be taken</code> 364 </p> 365 366 367 <h3> Language </h3> 368 369 <h4 id="attach">Attach</h4> 370 <p> The possibilities for creating errors when using 371 <code>attach</code> are numerous. Avoid it.</p> 372 <h4 id="functionlanguage">Functions</h4> 373 <p> Errors should be raised using <code>stop()</code>.</p> 374 <h4 id="object">Objects and Methods</h4> 375 <p> The S language has two object systems, S3 and S4, both of which 376 are available in R. S3 methods are more interactive and flexible, 377 whereas S4 methods are more formal and rigorous. (For an illustration 378 of the two systems, see Thomas Lumley's 379 "Programmer's Niche: A Simple 380 Class, in S3 and S4" in R News 4/1, 2004, pgs. 33 - 36: 381 <a href="https://cran.r-project.org/doc/Rnews/Rnews_2004-1.pdf"> 382 https://cran.r-project.org/doc/Rnews/Rnews_2004-1.pdf</a>.) 383 </p> 384 <p>Use S3 objects and methods unless there is a strong reason to use 385 S4 objects or methods. A primary justification for an S4 object 386 would be to use objects directly in C++ code. A primary 387 justification for an S4 generic/method would be to dispatch on two 388 arguments. 389 </p> 390 <p>Avoid mixing S3 and S4: S4 methods ignore S3 inheritance and 391 vice-versa. 392 </p> 393 394 395 <h3>Exceptions</h3> 396 397 <p> 398 The coding conventions described above should be followed, unless 399 there is good reason to do otherwise. Exceptions include legacy 400 code and modifying third-party code. 401 </p> 402 403 404 <h3>Parting Words</h3> 405 406 <p> 407 Use common sense and BE CONSISTENT. 408 </p> 409 <p> 410 If you are editing code, take a few minutes to look at the code around 411 you and determine its style. If others use spaces around their 412 <code>if </code> 413 clauses, you should, too. If their comments have little boxes of stars 414 around them, make your comments have little boxes of stars around them, 415 too. 416 </p> 417 <p> 418 The point of having style guidelines is to have a common vocabulary of 419 coding so people can concentrate on <em>what</em> you are saying, 420 rather than on <em>how</em> you are saying it. We present global style 421 rules here so people 422 know the vocabulary. But local style is also important. If code you add 423 to a file looks drastically different from the existing code around it, 424 the discontinuity will throw readers out of their rhythm when they go to 425 read it. Try to avoid this. 426 </p> 427 428 <p> 429 OK, enough writing about writing code; the code itself is much more 430 interesting. Have fun! 431 </p> 432 433 434 <h3>References</h3> 435 436 <p> 437 <a href="http://www.maths.lth.se/help/R/RCC/"> 438 http://www.maths.lth.se/help/R/RCC/</a> - R Coding Conventions 439 </p> 440 <p> 441 <a href="http://ess.r-project.org/">http://ess.r-project.org/</a> - For 442 emacs users. This runs R in your emacs and has an emacs mode. 443 </p> 444 445 </body> 446 447 </html> 448
