To support existing data analysis pipelines that use the Stata software, SESYNC has purchased a Stata license and created a dedicated virtual machine for remote use by affiliated researchers. This quick start guide explains the essential steps for evaluating Stata commands over SSH or from SESYNC’s RStudio server.
Stata is installed on a Linux server with address “stata.sesync.org”, that
affiliated researchers can access through a terminal/console using SSH. For
example, a user logged into our RStudio,
Jupyter, or gateway servers can open a terminal and execute
ssh stata.sesync.org to reach the server with the Stata software.
The first time you ssh to this or any server, you will be asked to confirm the
remote machine’s identity. Do this even if you only plan to use the RStata package.
<USERNAME>@rstudio03$ ssh stata.sesync.org
The authenticity of host 'stata.sesync.org (192.168.***.***)' can't be established.
ED25519 key fingerprint is SHA256:*******************************************.
Are you sure you want to continue connecting (yes/no)? yes
Warning: Permanently added 'stata.sesync.org,192.168.***.***' (ED25519) to the list of known hosts.
Welcome to Ubuntu 16.04.6 LTS (GNU/Linux 4.4.0-143-generic x86_64)
* Documentation: https://help.ubuntu.com
* Management: https://landscape.canonical.com
* Support: https://ubuntu.com/advantage
Last login: Thu Apr 4 10:50:22 2019 from 192.168.***.***
<USERNAME>@stata00:~$
With your terminal now connected to this server, launch the console version of
Stata for interactive use by calling stata. As on all our compute platforms,
your research data directory is accessible from the Stata console.
<USERNAME>@stata00:~$ stata
___ ____ ____ ____ ____ (R)
/__ / ____/ / ____/
___/ / /___/ / /___/ 15.1 Copyright 1985-2017 StataCorp LLC
Statistics/Data Analysis StataCorp
4905 Lakeway Drive
College Station, Texas 77845 USA
800-STATA-PC http://www.stata.com
979-696-4600 stata@stata.com
979-696-4601 (fax)
Single-user Stata perpetual license:
Serial number: ************
Licensed to: University of Maryland-SESYNC
1 Park Pl Suite 300 Annapolis MD 21401
Notes:
1. Unicode is supported; see help unicode_advice.
. exit
<USERNAME>@stata00:~$
The RStata package facilitates using R to send input and receive output from Stata. The package documentation describes communicating with Stata software installed on the same machine as R, but is conveniently written to also allow communication with a remote installation.
For the time being, you need a patched version of the RStata package in order to use it sucessfully on SESYNC’s servers. Use devtools to install the RStata package:
library(devtools)
install_github('lbraglia/RStata', ref = github_pull(8))
The package requires configuration of two options: the “path” to the Stata executable and the Stata version number. From our RStudio server, these options should be set as follows.
library(RStata)
options(RStata.StataPath='ssh -q stata.sesync.org /usr/local/stata15/stata')
options(RStata.StataVersion=15)
To confirm that the options are correctly set, use the stata function
to send the one-line command requesting Stata to print its version:
stata('version')
The stata function alternatively accepts the path to a Stata “.do”
file, which will be much more efficient than running multiple
one-liners. The ?stata help describes how to use data.in and
data.out arguments to handle data transmission, but we strongly
discourage their use. Instead, have both R and Stata read and write data
to a location on our network file system, accessible at files.sesync.org.
Both servers have access to the same data directories.
Here is a simple “.do” file that demonstrates reading and writing to your
project’s path <PROJECT>-data on the network file system.
use /nfs/<PROJECT>-data/training/census
tabulate TRACT
collapse RENTER_OCC OWNER_OCC MED_AGE
save /nfs/<PROJECT>-data/training/census_summary, replace
Assuming this file is saved as “example.do” in the working directory,
you can execute it from R using the stata function.
stata('example.do')
The haven package handles reading and writing Stata’s native .dta
files, for both recent and older versions of the Stata software. A
complete pipeline making use of R and Stata could look like the
following:
library(RStata)
library(haven)
options(RStata.StataPath='ssh -q stata.sesync.org /usr/local/stata15/stata')
options(RStata.StataVersion=15)
# read data into R
to_stata <- read.csv('/nfs/<PROJECT>-data/training/census.csv')
# process the to_stata data.frame in R
# ...
# save data to network file system
write_dta(to_stata, '/nfs/<PROJECT>-data/training/census.dta')
# continue processing in Stata
stata('example.do')
# read result back into R
from_stata <- read_dta('/nfs/<PROJECT>-data/training/census_summary.dta')
head(from_stata)