Skip to content

rFunctions.md

Latest commit

 

History

History
222 lines (169 loc) · 4.4 KB

rFunctions.md

File metadata and controls

222 lines (169 loc) · 4.4 KB

rHelp

Useful R function with quick syntax references.

These are some useful R functions for wrangling data. The functions include: aggregate, apply, bash engine, c (concatenate), colnames, ddply (for row counts), drop_na, filter, flags, grepl, group_by, gsub, ifelse, is.na, join, left_join, paste, rbind, read_excel, save, select, separate, separate_rows, sort, spread, substr, summarise, t.test, unique, write.csv, ...

aggregate: Sum samples within a group

df <- aggregate(df$varToFunctionalize, by df["varToGroup"], FUN = mean)

apply: anonymous, user defined or other functions

apply(df[1:3], 2, function(x) (min(x) + max(x)) / 2)             # 2 is for col, 1 is by row

midrange <- function(x) (min(x) + max(x)) / 2

apply(df[1:3], 2, midrange)

apply(df[1:3], 2, summary)                                       # this forms a matrix

sapply(df, function(x) round(coef(lm(df[1] ~ x, data = df)), 3)  # table of coefficients for linear regressions on variables in col2,3 against df col1

by: higher order function (function on function)

by(df[1:3], df$CategoricalVariable, summary)

bash in RStudio

bash engine='bash'

colnames: Slick column rename

colnames(df)[names(df) == "varName"] <- "newVarName"

ddply:: Summarize N of Variable Values

rowCountForVar <- plyr::ddply(df, .(df$groupingVar, df$varToCount), nrow)

drop_na: Remove rows which have NAs in one (key) variable

dfRmNa <- drop_na(df, "Var")

drop na for cols

AE <- read_csv("ae.csv")

DT_AE <- as.data.table(AE)

CONDENSED_AE <- DT_AE[,which(unlist(lapply(DT_AE, function(x)!all(is.na(x))))),with=F]

filter Variable on Value

df <- filter(df, df$\`Variable Name\` == "Value")

df <- filter(df, !grepl('ID1|ID2|ID3', idVar))

d.filt <- filter(d, d$Gene == "GoI")

t.test(d.filt$IC50~d.filt$Mutated)

Consolidate Variable Values

df$`Consolidated Vars` <- ifelse((df$"Var to Consolidate" == "high") | (df$"Var to Consolidate" == "very high"), "Responding Patient", df$"Var to Consolidate")

grep: Remove data

df <- filter(df, !grepl('ID1|ID2|ID3',idVar))

group_by: Summarise with group_by

df <- df %>% 
  group_by(groupVar) %>% 
  summarise_each(funs(min(resultVar, na.rm = TRUE)))

gsub: Remove commas, AKA Replace commas with nothing

df$`Some Variable` <- gsub(",", "", df$`Some Variable`)

ifelse, is.na: Flag samples without value for variable

df$`newVariable` <- ifelse(is.na(df$variable), FALSE, TRUE)

join, left_join: Join two dataframes (ie more columns)

df <- left_join(df, altDf, by = c("Var1" = "Var1", "Var2" = "Var2"))

paste: Create new variable

df$newVar <- paste("Year", df$year, "Month", df$month, "Day", df$`The Day`)

rbind: Adding dataframes together (ie more rows)

dfBound <- rbind(df1, df2)

read_excel: Read in Excel file

df <- read_excel("/Path/to/file.xlsx")

save

save(df, file = "df.RData")

select: Parse out columns of interest

dfSvelte <- select(dfSvelte, 1:42,45)
df <- df[,c("Col 1", "Col 2", "Col 3")]

separate: Splitting a variable column

df <- separate(df, `Var Name`, c("Token 1", "Token 2", "Token 3"), sep = "_")

separate_rows: Separate delimited in several columns into separate rows and fill other variables

dfSplit <- separate_rows(df, delimVar1, delimVar2, delimVar3, delimVar4, sep =  ";", convert = FALSE)

sort: UNIX-like (not R, use bash engine) by column (k, man sort)

sort -k1,1 /Users/jlong/file.txt > /Users/jlong/file.sorted.txt

spread: Transposing data

dfWide <- dfLong %>% 
  spread(`Row Name Var`, `Result To Spread`, convert = TRUE)

substr: Cut up regular values (by fixed width of string)

df$`The Var` <- substr(as.character(df$`The Var`), index1, index2)

summarise: aggregates

anotherDf <- df %>%
  group_by(groupVar) %>%
  summarise(df = mean(result))

table: create a table for the counts of a factor

table(df[1])

t.test: for Gene Mutation of Interest given IC50 scores for cell lines

d.filt <- filter(d, d$Gene == "GoI")
  t.test(d.filt$IC50~d.filt$Mutated)

unique: Remove duplicates

df <- unique(df)

str_to_upper: convert to uppercase

stringr::str_to_upper("anonymous function")

write.csv: Write data

write.csv(df, "Filename.csv", row.names=FALSE, na="")