dismisc: misc R functions

Guidance

In my projects I always have an init.R script that I run before all others. It varies based on project but typically looks something like this:

library(magrittr) # for %<>% and %$%
library(dplyr) # the main workhorse
library(readr)
library(stringr)
library(reshape2)
library(ggplot2)
# ... etc


library(dismisc) # this package

.rt_scratch <- "/path/to/scratch/directory/"
.rt_proj <- '"/path/to/proj/directory/" 
.rt_data <- '"/path/to/data/directory/" 
.rt_scripts <- '"/path/to/scripts/directory/" 

Functions in dismisc

setrt() and setwdrt()

Note: these are obsolete though kept for running old code. It's better to use hidden objects, like .rt_scratch <- "/path/to/scratch/directory/" etc.

Misc functions

• tlag, tlead - taken from https://github.com/matthieugomez/statar, lead/lag by a time variable

• tab - as tab from stata, but only works for vectors

• su_up - as su from stata, but only works for vectors

• clear_all_labels(df) - removes all labels from the columns of the dataframe df (there may be a better way)

• get_non_numerics(x) - given a vector returns the subvector that will go to NA when coerced

• download_them_all() - downloads files in a list of links

• dismisc_extrapolate() - extrapolate in a grouped dplyr::mutate using stats::splines or lm. (Special name to avoid conflicts.)

• mdb_get() - a simplified version of Hmisc::mdb.get.

• parse_censored_data_edfacts() - parses a particular time of numerically censored data. (Special name to avoid conflicts; named this way because edfacts data has this structure, although I've seen it elsewhere too.)

Possible replacements?

• get_non_numerics(). Maybe just do c("1","3","h","i","32","32.32") %>% .[is.na(as.numeric(.))] -- It's almost as much typing.

Functions removed when I moved from library(disalvo) to library(dismisc), with comments

These comments might help people who are trying to figure out how to solve certain common data problems that sound like they need a new function but actually can be solved easily with dplyr etc.

• db_init(). Don't make a function to load libraries etc.; make a script init.R for each project instead. This way you can call setrt() in the init.R script.

• cols.to.char(). Instead, use df %<>% mutate_all(funs(as.character(.)))

• lager()/leader(). Instead use statar::tlag in a dplyr::mutate.

• print.db.table() and other xtable hacks. Instead use R.rsp with a latex template file. More details on this in a future version of dismisc.

• empty.string.to.NA(). I used to use this a lot so that foreign::write.dta() worked. Instead use df %<>% mutate_all(funs(ifelse(.=="",NA,.))). Proof:

df <- data_frame(x = c("a","b",""), y = 1:3)
df %>% mutate_all(funs(ifelse(.=="",NA,.)))

• convertsp2.names and convert_2.names. Just do names(df) %<>% str_replace_all(pattern = "[ ]", replacement = ".") etc.