rmsfuns

Nico Katzke

2020-05-05

Introduction

This package contains several helper functions for use in data manipulation, folder creation and viewing purposes. See examples of such functions below.

build_path

This function builds the entire folder path provided by the user. If the path does not exist, it builds it without error. It is effectively a user-friendly wrapper to the base function dir.create.

library(rmsfuns)
build_path("C:/Temp/data")

Can also be used to build a vector of paths:

library(rmsfuns)
Path <- build_path(paste0("C:/Temp/data/", c("SubFolder1", "SubFolder2", "SubFolder3"))
print(Path)

ViewXL

This function makes it easy to quickly view any R object or dataframe in excel. A random file is created in R’s temporary folder location (see tempdir() to find your location). The excel file location can also be overridden using the FilePath command. IMPORTANT: if using a mac, set mac = TRUE in the command (equal to FALSE by default).

library(rmsfuns)
df <- data.frame(date = 
seq(as.Date("2012-01-01"),
as.Date("2015-08-18"),"day"), 
x = rnorm(1326, 10,2))

x <- ViewXL(df)
# After viewing in excel, delete the temporary file:
unlink(x)
# ViewXL(df, mac = TRUE) if using a mac

dateconverter

The dateconverter function makes it easy to create a date vector in R. It offers a simple wrapper using xts functionality to create a vector of dates between a given Start and End date, and then correcting for the chosen frequency transformation.

It can do the following transformations between given Start and End Dates:

alldays ; calendarEOM ; weekdays ; weekdayEOW ; weekdayEOM ; weekdayEOQ ; weekdayEOY

library(rmsfuns)
dates <- dateconverter(as.Date("2000-01-01"), as.Date("2017-01-01"), "alldays") 
dates <- dateconverter(as.Date("2000-01-01"), as.Date("2017-01-01"), "weekdays") 
dates <- dateconverter(as.Date("2000-01-01"), as.Date("2017-01-01"), "calendarEOM")
dates <- dateconverter(as.Date("2000-01-01"), as.Date("2017-01-01"), "weekdayEOW")
dates <- dateconverter(as.Date("2000-01-01"), as.Date("2017-01-01"), "weekdayEOM")
dates <- dateconverter(as.Date("2000-01-01"), as.Date("2017-01-01"), "weekdayEOQ")
dates <- dateconverter(as.Date("2000-01-01"), as.Date("2017-01-01"), "weekdayEOY")

PromptAsTime

To change R’s prompt to reflect the time, use the PromptAsTime function. This can be used as a simple means of timing long calculations without using sys.time() commands. This can be very useful if running, e.g., many functions overnight, and later viewing the time taken on multiple calculations.

To set the timer on, type:

PromptAsTime(TRUE)

The time for each command will now be shown in Rstudio’s prompt.

This is particularly useful for when you want to see, after running a code script in Rstudio, what the duration of each line was. E.g., run the following in your Rstudio console:

PromptAsTime(TRUE)
x <- 100
Sys.sleep(3) 
x*x
print(x)
PromptAsTime(FALSE)

You can then see in the prompt that the Sys.sleep(3) call lasted 3 seconds.

Safe_Return.portfolio

This function provides a safe alternative to do portfolio return calculations using PerformanceAnalytics::Return.portfolio.

See this gist for where the function fails: https://gist.github.com/Nicktz/a24ba1775d41aab85919c505ca1b9a0c

Problems with PerformanceAnalytics::Return.portfolio

  • For starters, the resulting xts starts on the second day - i.e. it produces weights and returns from the day after the initial weights are provided.

    • E.g. from the vignette it states: “Rebalancing periods can be thought of as taking effect immediately after the close of the bar. So, a March 31 rebalancing date will actually be in effect for April 1.”

    • In most applications this does not fit with my own personal workflow. If it fit yours, simply set lag_weights = FALSE.

  • Second (and most frustratingly) - the function is order dependent.

    • Thus it does not square stock X, Y and Z for R and weight inputs, but rather considers column orders.

    • This is extremely dangerous - and can easily cause unintended mistakes to enter your calculations.

To solve both the above issues - instead use this safer wrapper. The function works otherwise exactly like that of PerformanceAnalytics::Return.portfolio

load_pkg

This function loads a vector of packages into R, and installs the package if it has not yet been installed.

Packages <- c("xts", "dplyr")
load_pkg(Packages)