R Programming Best Practices Checklist

Galleries: R Markdown, Graphs, Shiny, R4CSR Tables, Flextable, Posit Tables

Base R

1. Good Programming Style for Beginners

• Keyboard shortcuts: 
• • <- from Alt +
  • %>% from Ctrl + Shift + m

  • Ctrl + Enter – Will run current line and jump to the next one, or run selected part without jumping further.

  • Alt + Enter – Allows running code without moving the cursor to the next line if you want to run one line of code multiple times without selecting it.
  • Ctrl + Alt + R to run whole script and
  • Ctrl + Alt + B/E combinations to run it from Beginning to the current line and from the current line to the End.
• R program names should have '.R' suffix, ex. mycode.R
• R variable names should not contain dot '.' and not the same name as R functions, ex. split_1
• Since variable names are case sensitive, it is best to rename them all to be lower case for easy or R programming
• To help standardize character variable conditions, apply TOLOWER() or TOUPPER() functions to lower or upper case values, names(aslx)= tolower(names(aslx))
• New R function names should be verbs
• Use '<-' instead of '=' for creating vectors, data frames and assignments, ex. x <- 10.  Left side of '<-' should be valid vector or data frame name and right side should be valid R function and syntax.
• Assignments to single variables in data frames may be referenced by [index position], taadmin[[6]] <- (!duplicated(taadmin$PT)) 
• Best to write R comments in the beginning of the line instead of middle or end, Single line comments are possible.
• To comment multiple lines of R syntax, one technique is to wrap it around an R function. The alternative within R Studio is to select the block of lines and select Control + Shift + C to insert # in front of each line.        
• skip <- function(x) {   
  
•   myquery <- ToothGrowth %>% # 1. source df
•   select(len, supp, dose) %>% # 2. select variables, - drop
•   filter(supp == 'VC') %>% # 3. subset records
•   mutate(dose2 = (dose*2)) %>% # 4. derive variables
•   arrange(supp, dose) # 5. sort records, desc()
•   }
• For small number of variables and records, create each variable vector and then combine them to create data frame 
• For large data files, it is best to load data into data frame, for csv files, missing values are stored as 'NA', for reading SAS datasets, missing values may be saved as ., '' or 'NA', best to double check missing values in data frames
• In general, create vectors and data frames instead of matrices
• select(dm, studyid, subjid, age) to include source data frame as first parameter instead of using default data frame
• Import only variables and records of interest from data files
• Subset data frames as soon as possible
• For each R function, always write '(' and ')' and then fill in parameters to not forget the ending ')' 
• For R functions that support multiple variables, separate variables by ','
• Create intermediate data frames instead of writing R multi-tasking
• Create test data frames with R functions before replacing data frames
• Create and test each R task component (filter, select, mutate, arrange, summarize) before assembling them using pipe %>%.  In general R task components can be in any order but best practices are to filter records, select variables, derive variables  and then arrange and summarize at the end.  In addition, group_by() should be before summarize() for by group summaries instead of outputting overall summaries.  Nesting R functions is an alternative to using pipe %>% but not ideal for more than 2 R task components.  For nesting R functions, it is best to indent each R component in new lines.  
• Leverage popular R package such as Tidyverse and DPLYR for data management operations such as joins.  
• Common non-by variables in joins are renamed to <name>.x and <name>.y.
• Install and load R packages before calling R functions
• Apply common R functions instead of applying other R functions for similar tasks
• Copy/paste/update working R example for tasks instead of trying to remember exact R function syntax
• Cearch for common R solutions to address common R errors
• if (missing(df)) stop("parameter df is required.") for confirming dataframe exists
• dim(df) check the number of variables and rows
• is.na(df) check if the data frame has missing values
• if (any(is.na(df))) print("missing values exist in at least one variable")
• In a function call, arguments can be specified by position, complete name, or partial name. Never specify by partial name and never mix by position and complete name, ex. mean(x, na.rm = TRUE)
• rnorm(10, 0.2, 0.3)

• The na.rm argument appears in many R functions. R uses NA to represent Not Available, or 

  missing values. In summary functions, best to set na.rm=TRUE to ignore missing values since this is not done by default, ex. mean(mydata,na.rm=TRUE).
• Put spaces around all binary operations, ex. x == y
• Put a space after comas for second parameter, ex. mtcars[1, ]
• Put a space before left parentheses, except in a function call, ex. if (grade == 5.5)
• No spacing around code in parenthesis or square brackets, ex. species["tiger", ]
• Indent code but not using tabs
• When new variables are created from variables that contain NA, then NA will be the new variable result unless you include and !is.na(df$col1) in your condition to convert NA to the label for false values.  In general, the check for NA values is not needed. 
• When transforming values, RECODE() works similar to Proc Format
• When writing conditions for creating new variables or subetting records, it best to use CASE_WHEN() for greater flexibility in creating multiple expressions and conditions
• Break long R code into several lines, ex. long_function_name <- function(arg1, arg2, arg3, arg4, long_argument_name1 = TRUE)
• Comments are useful to explain logic behind R code and placed before or on the same line as R code, ex. # Derive age
• Do NOT put more than one statement (command) per line. 
• Do NOT use semicolon as termination of the command, ex. x <- 1
• x <- x + 1
• Cleanup objects and workspace before submitting new R program, ex. remove(list=ls()) for objects and rm(list = setdiff(ls(), "data")) for data frames
• Understand your data before performing analysis, ex. summary(data)
• Understand your data by groups before performing analysis, ex. aggregate(y ~ group, data, mean)  

For graphs, it is best to transpose data to be long and thin, ex. 

• data_long <- as.data.frame( # Convert data frame to long format
• pivot_longer(data = data,
• cols = c("x1", "x2", "x3", "y")))
• head(data_long)

1. Read and Write Data
   1. Text, Excel, SAS
2. Create, and Select Variables and Observations
   1. Selecting Variables
   2. Selecting Observations
   3. Converting Data Structure
   4. Data Conversion Functions
3. Combining Variables into Data Frames for Data Management
   1. Transforming Variables
   2. Conditional Transformations
   3. Logical Operators
   4. Conditional Transformation to Assign Missing Values
   5. Multiple Conditional Transformations
   6. Renaming Variables and Observations
   7. Recoding Variables
   8. Keeping and Dropping Variables
   9. By Group Processing
   10. Concatenating Observations
   11. Joining Data
   12. Summarize Data
   13. Reshaping Variables and Observations
   14. Sorting Data
4. Value Labels and Formats
5. Variable Labels
6. Graphics
7. Analysis

Top 18 R Progrmaming Interview Questions