Functions for Robust Variable Handling

Variables are the fundamental parts of each programming language but handling them efficiently might be frustrating for programmers. This package contains some functions to help user (especially data explorers) to make more sense of their variables and take the most out of variables and hardware resources. These functions are written, collected and crafted over 7 years of experience in statistical data analysis on high-dimensional data, and for each of them there was a need. Functions in this package are suppose to be efficient and easy to use, hence they will be frequently updated to make them more convenient.


News

varhandle 2.0.3

Changes to existing functions

  • rm.all.but()
    • Fixed a bug that was only expecting regular expression.
  • pin.na()
    • Added the possibility to define the missingness character or value (can be more than one). So now user can define what should be concidered as missing value (e.g na.value = c(NA, " ", "."))
  • inspect.na()
    • The added feature to pin.na() was also added to this function.
    • Fixed the issue that the code was breaking if the given matrix didn't have columns names.
  • check.numeric()
    • Fixed a bug that was returning TRUE when a numeric vector with some continuous was provided along with the flag only.integer = TRUE. Now the function checks the entire vector when it is of class numeric or integer and the flag only.integer is turned on.
  • var.info()
    • Now support showing the dimention of object with class "Matrix" from a package with the same name in the detail column of the output.
    • Now support showing the length of lists in the detail column of the output.

varhandle 2.0.2

Changes to existing functions

  • var.info()
    • Added progressbar and an argument to turn it on or off. Default is on.
    • Fixed a bug that was returning a warning when user was providing more than one variable name.
  • unfactor()
    • Fixed a bug that when a vector was fed, the function was retuning a warning. (now compatible with _R_CHECK_LENGTH_1_CONDITION_)
  • rm.all.but()
    • Added the ability to auto detect and handle regular expression alone or in combination with varibale names, so that it is more convenient for user to keep variables based on regular expression as well.

varhandle 2.0.1

Changes to existing functions

  • pin.na()
    • Change the type of output to data.frame to make it easier to access via $.
    • Now returns NULL in case it does not find any NA. This change has been done to make it easier to combine it with is.NULL()
  • check.numeric()
    • The rm.na argument has changed to na.rm in order to make it similar to the convention that other packages and functions are using.
    • The function now detects "-.2", "3.", "" and NA as numbers as well.
    • The default value of argument na.rm has changed to FALSE in order to take NAs into account.
    • An option added to ignore leading and tailing whitespace characters from items in vector before assessing if they can be converted to numeric.
  • rm.all.but()
    • Added the ability to call garbage collection if the size of the removed variables exceed the new parameter gc_limit.
    • Added a new parameter keep_functions to automatically exclude all functions from being removed.
  • var.info()
    • Now can handle matrix-like objects with multiple classes.

New functions

  • inspect.na(): This function is calls pin.na() and produce a human readable data.frame of NA status of columns in addition to a barplot and/or histogram.

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.

install.packages("varhandle")

2.0.4 by Mehrad Mahmoudian, 2 months ago


https://bitbucket.org/mehrad_mahmoudian/varhandle


Report a bug at https://bitbucket.org/mehrad_mahmoudian/varhandle/issues


Browse source code at https://github.com/cran/varhandle


Authors: Mehrad Mahmoudian [aut, cre]


Documentation:   PDF Manual  


GPL (>= 2) license


Imports utils, graphics


See at CRAN