# Introduction

The rccmisc package contains functions either required by other Swedish Regional Cancer Center packages (see: https://bitbucket.org/cancercentrum/rcc2) or standalone functions outside the scope of these packages.

To get an overview over all exported functions, please use: help(package = "rccmisc").

Functions of the package might also be categorised under the themes below. Please see ?function for function names printed as function!

# INCA infrastructure

The rcc suite of packages are intended to work both locally and on INCA. An assertion function is.inca might be used to check if a script is called from INCA or not.

Sometimes you want to use a package not currently installed on the INCA R server. Then use make_r_script to export your script together with all needed functiuons from the package required.

# Miscellaneous help functions

Large data sets from INCA might contain several hundreds of variables. It is often not possible to remember all available variable names wherefore functions findvar, findvar_anywhere, findvar_fun and findvar_in_df might come handy.

Another problem regarding INCA variable names is that names are case sensitives while many statisticians prefer working with lower case variable names only. A common solution to this problem is to start R-scripts with: names(df) <- tolower(names(df)). This is however dangerous since more than one variable name might get the same lower case representation (this is a fact for some poorly designed variable names). Use function lownames instead.

A related function is change_col_name that will ensure that a name change will not result in a data frame with more than one column of the same name (which is possible but potentialy dangerous in R).

Missing values might be coded differently for different variables (such as “99” or “999” etc). Use specify_missing to handle missing values correctly.

Base R contains some handy functions for “summarising in parallel” such as pmin and pmax. We here add psum to calculate sums in a similair way.

Base R also have a function called cut used for categorisation of numeric data. Numeric vectors are handled by S3-method cut.default that does not take actual values into consideration. Our method cut.integer makes a nicer output when all numeric values happen to be integers.

The ifelse function of base R is known to be dangerous (sometimes converting its outcome in unpredictable ways). safe_ifelse is a safer alternative!

It migt sometimes be useful to know the “width” of an interval. This might be done manually using range but the new function width just makes it a little simpler.

# Lowlevel miscellaneous functions

INCA data is sometimes converted between different formats. A numeric variable might sometimes end up being a character or factor although its contant is still numeric in nature. Functions such as is.scalar_in, is.scalar_in01, is.wholenumber, is_numeric and as_numeric could come handy in this situation.

# Text handling

R is known for sometimes unnecesarly treating characters as factor. INCA is sometimes known for the slight opposite. Doctors name for example might be missspelled or sometimes prefixed by title etcetera. Funtions best_match and clean_text might be useful in this situation.

# Internal help functions

Some functions will probably not be that interesting outside the world of package development. The package does however also contain some internal help functions for other RCC packages, such as create_s3_method and create_s3_print.