Draft sas and r and sas (may, 2018 asa meeting)

SAS AND R AND SAS
Barry DeCicco
Ann Arbor Chapter, American Statistical
Association
May 22, 2018

CONTENTS
Calling R from SAS.
Calling SAS from R.
References.

CALLING R FROM SAS
USING SAS/IML

WHY USE R IF YOU HAVE SAS?
You can use SAS for what you already know well, and use
R only as needed.
You might find that you don’t have the capacities in SAS
which you desire.
R has vast and rapidly growing capacities – over 10,000
packages.

BASICS OF R
R must be installed on your computer. Download from
CRAN: https://cran.r-project.org
Necessary packages (modules) must be downloaded
from CRAN and installed: (select ‘Packages’ from the left-
hand menu).
You will need to set up SAS to use R.

SETTING UP SAS TO USE R
You need to edit the file ‘SASv9.cfg’, which controls
some configurations (run ‘proc options option=config;’
to find it).
Make a copy, and then edit the original, adding these
lines:
1. -RLANG
2. -config "C:Program
FilesSASHomeSASFoundation9.4nlsensasv9.cfg“[replace
with folder name for where this file is on your computer]
3. -SET R_HOME "C:Program FilesRR-3.5.0“ [replace with folder
name for where R lives on your computer]

HOW TO SUBMIT R CODE IN SAS
When running analyses/data manipulations in R, there
are things which need to be done:
Setting directories as needed, to pull/save data, output
and graph files.
Loading (already installed!) necessary packages for
use (e.g., ‘ggplot2’ for complex graphs).
These will need to be done in the SAS code
submitted to R each time.

STARTING LINES
Use PROC Options to enable SAS to use R:
proc options option=rlang; run;
Start IML (this has to be done through SAS/IML):
Proc IML;
Export SAS datasets to R as needed:
Call ExportDataSetToR(SAS-data-set, RDataFrame);
Let SAS know to start R:
submit / R;

ENDING LINES
Retrieve data sets from R as needed:
 CALL IMPORTDATASETFROMR (SAS-data-set, RExpr) ;
Let SAS know to end sending commands to R:
endsubmit;
Quit IML (if no further IML use is needed):
quit;

R LANGUAGE COMMANDS
Set the working directory (location where R retrieves and
stores files), while storing the previous location in ‘old_wd’:
 old_wd<-setwd("C:/Users/Barry/Desktop")
Resetting the working directory back to what it was:
setwd(old_wd);
# this is a comment

GOING THROUGH THE LOG -
PRELIMINARIES
 /* Start R options/ */
 proc options option=rlang;
 run;
 proc iml;
 title "Statistic in R (integration with SAS)";
 run ExportDataSetToR("WORK.RCBD_TEST", "RCBD_IN_R");
 submit / R;

GOING THROUGH THE LOG –
SUBMITTING COMMANDS TO R
 #_______Beginning of R code_________________
 # Change the working directory, saving the old one in ‘old_wd’
 old_wd<-setwd("C:/Users/Barry/Desktop")
 # Load the ‘qicharts’ module (package), so that it can be used.
 library(qicharts)

 # Create a data frame (data set)
 d2 <- data.frame(y = rnorm(144, 12, 3),
 expand.grid(x = seq.Date(as.Date('2014-1-1'),
 by = 'week',
 length.out = 24),
 g1 = LETTERS[1:3],
 g2 = letters[1:2]))
 # Introduce a shift in process performance
 d2$y[132:144] <- d2$y[132:144] * 3

 # Run a chart.
 trc(y ~ x | g1 + g2, data = d2, main = 'Trellis run chart', chart = 'i')
 # Direct the graphics output to a file, then close it.
 dev.copy(pdf,'myplot.pdf',width=6,height=4 )
 dev.off()
 # Reset the working directory
 setwd(old_wd)

WRAPPING UP
 #_______End of R code_________________
 endsubmit; # Any error messages from R will now show up
 * Bring the data set from R to SAS;
 CALL IMPORTDATASETFROMR ("work.d2", "d2") ;
 proc print data=d2;
 run;

CALLING SAS FROM R
USING KNITR AND
MARKDOWN

WHAT IS KNITR?
knitr is an R package which combines code,
comments, text and results into one
document.
You write a document with text and code
chunks.
knitr will run the code chunks and incorporate
the results, knitting them into a single
document.
This document can be in Word, HTML or PDF
(the latter requires outside software).

THREE OUTPUT OPTIONS
HTML – easiest to start
with.
Word.
PDF – this requires
external software to
finish.
I will use HTML

STARTING THROUGH THE DOCUMENT
Knitr will set up a
standard starting
document
prepopulated with
some items.
Title information.
Initial setup.
Explantion

CODE CHUNKS AND SETTING UP SAS
 Gray items are
‘code chunks’,
where there is code
to be executed.
 They start and end
with ```
 This sets up SAS in
this code chunk. It
tells R to use SAS,
and where SAS is.

LIMITATIONS
Each code chunk runs a separate batch job in
SAS. There is no hand-off of WORK. data sets
between SAS code chunks.
To hand off data sets (between SAS and SAS,
or SAS and R), save them as permanent files.
SAS HTML output plays well with knitr’s HTML
markdown. For other types of output, other
packages are needed.
Slashes/back-slashes in path names need to
be managed.

SET UP
Let R know that it’s using sas (the ‘SAS’ or
‘SASHTML’ engine).
Tell R where the SAS.exe file is.
Set any SAS options.
These can by done either by chunk, or overall
(for now, overall).

SET UP
There are at least 20 ‘engines’ available for
knitr, meaning 20 languages.
For SAS, there is the ‘sas’ and ‘sashtml’
engines.
For this presentation, the ‘sashtml’ engine will
be used.

SET UP
Once the set up
has been done
overall, then SAS
commands can
be done.

REFERENCES – SAS
 SAS/IML(R) 9.3 User's Guide ‘The RLANG System Option’
http://support.sas.com/documentation/cdl/en/imlug/64248/HTM
L/default/viewer.htm#imlug_r_sect003.htm
 SAS/IML(R) 9.3 User's Guide‘Submit R Statements’
L/default/viewer.htm#imlug_r_sect004.htm
 SAS/IML(R) 12.1 User's Guide: EXPORTDATASETTOR Call
L/default/viewer.htm#imlug_langref_sect114.htm
 SAS/IML(R) 12.1 User's Guide: IMPORTDATASETFROMR Call
L/default/viewer.htm#imlug_langref_sect182.htm

REFERENCES – SAS (CON.)
 ‘Run R code inside SAS easily’ SAS/IML(R) 9.3 User's Guide
‘EXPORTDATASETTOR Call’
http://support.sas.com/documentation/cdl/en/imlug/65547/HTML/default
/viewer.htm#imlug_langref_sect114.htm
 SAS Blogs: Video: Calling R from the SAS/IML Language
https://blogs.sas.com/content/iml/2011/10/31/video-calling-r-from-the-
sasiml-language.html

REFERENCES - R
 ‘SAS Using R Markdown (Windows)’
https://www.ssc.wisc.edu/~hemken/SASworkshops/M
arkdown/SASmarkdown.html#html-or-pdf-document-
as-a-final-target
 The raw markdown version of the above page:
arkdown/SASmarkdown.rmd
 ‘Tables and Graphs - Using the sashtml Engine’
arkdown/SAShtmlengine.html
 Dynamic Documents with R and Knitr (by Yihui Xie)
https://yihui.name/knitr/

Draft sas and r and sas (may, 2018 asa meeting)

More Related Content

What's hot (20)

Similar to Draft sas and r and sas (may, 2018 asa meeting) (20)

More from Barry DeCicco (6)

Recently uploaded (20)

Draft sas and r and sas (may, 2018 asa meeting)