Aleksander Rutkowski
2016-08-17
The available Stata commands -- very simple API:
r-- allows for the execution of a line of arbitrary R expressionsstartr-- starts a local R "server"stopr-- stops a local R "server"tor-- exports the current dataset (sitting in Stata memory) into R (tor= to R) to a data.frame named StataDatafromr-- import the StataData data.frame from R into Stata
Either automatically (from Stata) or manually (e.g. if Stata cannot download via HTTPS due to a web proxy configuration resulting in Stata error number 672 -- see http://www.stata.com/support/faqs/web/common-connection-error-messages/, paragraph on r(672), point 2):
Stata code:
cap mkdir "`c(sysdir_personal)'", public
loc gh "https://raw.githubusercontent.com/alekrutkowski/StataRLink/master/"
loc F1 r.ado startr.ado stopr.ado tor.ado fromr.ado check_data_consistency.ado
loc F2 StataRLink_server.R dicen.ado filewrap.ado waitforfile.ado
foreach f in `F1' `F2' {
di "Dealing with file `f'"
copy "`gh'`f'" "`c(sysdir_personal)'`f'", text public
}
If you re-install StataRLink, add the word replace at the end of the
penultimate line (the one inside the foreach loop,
which starts with copy), that should look like this:
copy "`gh'`f'" "`c(sysdir_personal)'`f'", text public replace
If you un-install StataRLink, replace the
penultimate line (the one inside the foreach loop,
which starts with copy) with the following code:
rm "`c(sysdir_personal)'`f'"
Download https://github.com/alekrutkowski/StataRLink/archive/master.zip
and open (unzip) it. Then, copy the unzipped contents
(the contents of the StataRLink-master subdirectory) into your "personal"
Stata directory. You can find out where your "personal" Stata directory is by
typing in Stata command personal. This directory may actually not exist yet --
you may have to create it first (you can do it from Stata:
mkdir "`c(sysdir_personal)'", public).
Important! Before starting the R "server" with command startr,
you have to tell Stata where to find it. Put the nested quotes
inside the rscript_path global macro if the path contains
one or more spaces:
E.g., on Windows:
global rscript_path `""C:\Program Files\R\R-3.2.0\bin\x64\Rscript.exe""'
or for a more complicated case (R as a virtualised app):
global rscript_path `""C:\Program Files (x86)\Microsoft Application Virtualization Client\sfttray.exe" /launch "Rscript [V] 3.24bd001""'
E.g., on Linux (Ubuntu) it may be:
global rscript_path "/usr/lib/R/bin/Rscript"
// Start R "server":
startr
R "server" started successfully
// Open an example dataset available in Stata (built-in):
sysuse auto, clear
(1978 Automobile Data)
// Keep only some columns (variables) for clarity:
keep make price mpg rep78
// Command "tor" = "to R" -- export data to R;
// by default, the dataset is put into R object called "StataData":
tor
Exporting data from Stata...
Importing data into R object 'StataData'...
------------------------------- R output start ---------------------------------
> str(StataData)
'data.frame': 74 obs. of 4 variables:
$ make : Factor w/ 74 levels "AMC Concord",..: 1 2 3 7 8 9 10 11 12 13 ...
$ price: int 4099 4749 3799 4816 7827 5788 4453 5189 10372 4082 ...
$ mpg : int 22 17 22 20 15 18 26 20 16 19 ...
$ rep78: int 3 3 NA 3 4 3 NA 3 3 3 ...
-------------------------------- R output end ----------------------------------
// The "r" command executes R functions interactively:
r StataData <- within(StataData, new <- price/100)
------------------------------- R output start ---------------------------------
> StataData <- within(StataData, new <- price/100)
-------------------------------- R output end ----------------------------------
// Import modified data ("StataData" object) back into Stata:
fromr, clear
Exporting 'StataData' from R...
Importing data into Stata...
(5 vars, 74 obs)
// Have a look at the top of the data:
list in 1/5, clean
make price mpg rep78 new
1. AMC Concord 4099 22 3 40.99
2. AMC Pacer 4749 17 3 47.49
3. AMC Spirit 3799 22 . 37.99
4. Buick Century 4816 20 3 48.16
5. Buick Electra 7827 15 4 78.27
// Do some manipulations again in R:
r StataData <- within(StataData, new2 <- rep78/100)
------------------------------- R output start ---------------------------------
> StataData <- within(StataData, new2 <- rep78/100)
-------------------------------- R output end ----------------------------------
// Import the latest version of the "StataData" object:
fromr, clear
Exporting 'StataData' from R...
Importing data into Stata...
(6 vars, 74 obs)
// Have a look at the top of the data:
list in 1/5, clean
make price mpg rep78 new new2
1. AMC Concord 4099 22 3 40.99 .03
2. AMC Pacer 4749 17 3 47.49 .03
3. AMC Spirit 3799 22 . 37.99 .
4. Buick Century 4816 20 3 48.16 .03
5. Buick Electra 7827 15 4 78.27 .04
// You can export observation subsets with "if" and "in"
// (put option "replace" to replace the current "StataData")
tor in 1/3, replace
Exporting data from Stata...
Importing data into R object 'StataData'...
------------------------------- R output start ---------------------------------
> str(StataData)
'data.frame': 3 obs. of 6 variables:
$ make : Factor w/ 3 levels "AMC Concord",..: 1 2 3
$ price: int 4099 4749 3799
$ mpg : int 22 17 22
$ rep78: int 3 3 NA
$ new : num 41 47.5 38
$ new2 : num 0.03 0.03 NA
-------------------------------- R output end ----------------------------------
// See how the data now looks inside R:
r StataData
------------------------------- R output start ---------------------------------
> StataData
make price mpg rep78 new new2
1 AMC Concord 4099 22 3 40.99 0.03
2 AMC Pacer 4749 17 3 47.49 0.03
3 AMC Spirit 3799 22 NA 37.99 NA
-------------------------------- R output end ----------------------------------
// You can also export column (variable) subsets:
tor m* in 1/3, replace
Exporting data from Stata...
Importing data into R object 'StataData'...
------------------------------- R output start ---------------------------------
> str(StataData)
'data.frame': 3 obs. of 2 variables:
$ make: Factor w/ 3 levels "AMC Concord",..: 1 2 3
$ mpg : int 22 17 22
-------------------------------- R output end ----------------------------------
// Again, see how the data now looks inside R:
r StataData
------------------------------- R output start ---------------------------------
> StataData
make mpg
1 AMC Concord 22
2 AMC Pacer 17
3 AMC Spirit 22
-------------------------------- R output end ----------------------------------
// You can also export to and import from a different object than
// the default "StataData"
tor m* in 1/3, name(myOtherName)
Exporting data from Stata...
Importing data into R object 'myOtherName'...
------------------------------- R output start ---------------------------------
> str(myOtherName)
'data.frame': 3 obs. of 2 variables:
$ make: Factor w/ 3 levels "AMC Concord",..: 1 2 3
$ mpg : int 22 17 22
-------------------------------- R output end ----------------------------------
// Escape the dollar character with a backslash to avoid
// interpreting it in Stata as a global macro
r myOtherName\$new_column <- 11:13
------------------------------- R output start ---------------------------------
> myOtherName$new_column <- 11:13
-------------------------------- R output end ----------------------------------
fromr, name(myOtherName) clear
Exporting 'myOtherName' from R...
Importing data into Stata...
(3 vars, 3 obs)
list, clean abbreviate(11)
make mpg new_column
1. AMC Concord 22 11
2. AMC Pacer 17 12
3. AMC Spirit 22 13
// Close the R "server":
stopr
The Stata command r returns the displayed R output in r-class Stata macros
plus a scalar indicating the number of lines returned. Type return list
after r ... to see it.
In addition, if the line of R code is executed with an error, r returns Stata error
number 675.