Nicktz

install.packages("rmarkdown", repos = "https://mran.revolutionanalytics.com/snapshot/2016-01-02")

Also see: https://github.com/fairtree/production.src/wiki/R-and-Revolution-R

Nicktz

See this post: http://stackoverflow.com/a/27027681/4198868

Nicktz

If e.g. a mutate wants to be done on multiple columns, we could use lapply together with lazyeval:

colfilter <- 
  lapply( ColumnNames, function(cols) lazyeval::interp(~Funcation(a), 
                                                       .values = list(a = as.name(cols))) )
df %>% mutate_( .dots = colfilter )

Nicktz

library(far)

data <- 
  data.frame(x = rnorm(30, 0, 1.5), 
             y = rnorm(30, 0, 1.5),
             z = rnorm(30, 0, 1.5))

y <- 
  orthonormalization(data,basis=FALSE, norm=TRUE)
# basis = TRUE squares columns.

Nicktz

# Simple:

select_(dataframe, .dots = VectorofNames)

Nicktz

As I answered here: http://stackoverflow.com/a/37939267/4198868

To remove all columns with only zeros:
    dfzeroremoved <- df %>% .[,colSums(. != 0) > 0]

To remove all columns with only NA:
    dfzeroremoved <- df %>% .[,colSums(!is.na(.)) > 0]

Nicktz

test <- function(x) {
y <- x^2
y
}

Nicktz

RdsFilesIdentical <- function(RdsLocation1, RdsLocation2) {
  
  library(fairtreeR)
  load.packages()
  
  Rds1 <- read_rds(RdsLocation1)
  Rds2 <- read_rds(RdsLocation2)
  
  identical(Rds1, Rds2)

Nicktz

dplyr allows the user to avoid loops. 

The loops are replaced by group_by() if data is well gathered into a tidy data frame.

E.g.:

  data <- data.frame(
  date = rep(c(1,2,3,4), each=25),
  Tickers = rep(c("A", "B", "C", "D")),
  Returns = rnorm(100),

Nicktz

Using, e.g., summarise_each, we can plug in ANY function and apply it to all groups!!

E.g., getting a boxplot table, by grouping according to a factor, and calculating the moments:
Data <- tbl_df ( Factors (str) | FactorValues (dbl) )

BoxMoments <-  Data %>% 
      group_by(date, Universe, Factors) %>% 
      summarise_each( funs( Min = min, Max = max, Mean = mean, Median = median, N = n(),
                          LowHinge = boxplot.stats(.)[[1]][2],  # Even other functions
                          UpHinge = boxplot.stats(.)[[1]][4]