Top
10 - Best SAS Tutorial Documents - Examples and Tips - A
Statistician's List
Daily sas code
snippets are posted in the site:
http://sascodes.blogspot.com
Keywords: SAS
macros, SAS SQL, SAS Table, SAS Data Step, ODS, Data Mining
Automation, SAS Tips, SAS Examples, SAS Reference for PROC
Statements, Analytics Automation
Section: SAS SQL:
SAS SQL has some special advantages; before you know that let
us understand how to do the same thing in standard SAS codes
and SAS SQL codes
A nice SAS SQL basics notes with examples
Top 10 SAS SQL examples:
http://www2.sas.com/proceedings/sugi29/042-29.pdf
A great simple SAS SQL teaching document
http://www.pauldickman.com/teaching/sas/proc_sql_code_20041026.pdf
A nice article about INTO Clause
http://www.ats.ucla.edu/stat/sas/library/nesug98/p193.pdf
Undocumented and Hard to find SQL Features
http://www2.sas.com/proceedings/sugi28/019-28.pdf
Part of this new book - Proc SQL: Beyond the Basics Using
SAS - By Kirk Paul Lafler where a great collection of codes
(To make sense you may benefit well buying the book - look at
SAS site for the book)
http://www.sas.com/samples/A58316
PROC SQL - Minimal Example Codes
- /*A good collection of PROC SQL Examples from the above
references*/
Section: SAS MACROS:
A great simple start to macro writing
Sugi 28: Nine steps to Start Writing Macros
A place of general collection of SAS starting points, SAS data
sets, Teaching kits with examples from different subjects is:
http://www.ats.ucla.edu/stat/sas/
A classic
macro example from "Nine Steps..." from above:
%macro
vacation(dsn=expenses,varlist=_all_);
proc sql
noprint;
Select
mean(RoomRate),
put(today(),mmddyy10.),
month(today())
into :average,:date,:mon
from &dsn;
%if &mon=6
or &mon =7 or &mon =8
%then
%do;
proc print
data=&dsn;
title
"Lowest Priced Hotels in the
%upcase(&dsn)
Data Set";
footnote
"On &date";
var &varlist;
where
RoomRate<=&average;
run;
%end;
%else
%do;
proc print
data=&dsn;
title "All
Room Information in the
%upcase(&dsn)
Data Set";
footnote
"On &date";
var &varlist;
run;
%end;
%mend;
options
symbolgen mprint mlogic;
%vacation(
)
========================================
Managing storage and recall of macros
========================================
Some not so commonly
known sas codes
ch_auto_open_date_2_f=input(ch_auto_open_date_1,yymmn10.);
- using do loop in a macro to read text based file
names and its contents as data:
STORAGE TECHNIQUES (SUGI Notes)
There are three ways to store macro programs for future use:
1) Store the source code in an external file and use this statement to
pull it into a SAS program:
%include(file-pathname);
2) Use the Autocall Facility to search predefined source libraries for
macro definitions. These libraries can consist of external
files or SAS catalogs.
3) Store the compiled macros. This does NOT save the SOURCE code! Always
maintain the source code separately. To store the compiled
macro :
libname dlib ‘library-path’;
options mstored sasmstore=dlib;
%macro mname(parameters) / store
des=’description’;
…
%mend;
To access this stored macro :
libname dlib ‘library-path’;
options mstored sasmstore=dlib;
%mname(parameters)
========================================
Exporting and
Importing Data
========================================
Importing data
from Excel: What if the data is access?
PROC
IMPORT
OUT= WORK.Howell
DATAFILE=
"E:\SPSS\Howell.xls"
DBMS=EXCEL2000 REPLACE;
GETNAMES=YES;
RUN;
PROC EXPORT DATA=SASUSER.CUST
OUTTABLE="CUSTOMERS"
DBMS=ACCESS
REPLACE;
DATABASE="C:\MYFILES\MYDATABASE.MDB";
RUN;
EXPORTING TO AN
ASCII (DELIMITED) FILE:
PROC EXPORT DATA=MYFILES.CLASS
OUTFILE="/MYFILES/CLASS"
DBMS=DLM;
DELIMITER=',';
RUN;
=====================================
Section:
Automating SAS outputs for summary measures in Data Mining:
Knowing output
dataset names in various PROCS in SAS STAT:
1. Use the
command outside of proc statements;
(before the specific proc statement) ODS TRACE ON;
(after that proc statement ) ODS TRACE OFF:
2. Run the
program and look at the log and find dataset name of output
files from any Procs; Check the sas output datasets to clearly
understand name of the file and the contents;
3. Now take
out those ODS Trace statments; introduce the following before
the PROC statment:
4. ODS listing
Name (file name from ODS trace)=New
name (it could have a specific path); /* storing it in
a particular path and will not be lost after the SAS run)
5. ODS listing close;
6. Now do the
proc contents to know the column
names and attributes.
7. In the next
run use the sas dataset name as a dataset to access, to pull
out your elements selectively, and to format the outputs to
your needs;
This is extremely
useful for automation and for large scale data mining
problems.
For some more
details look into
http://jeff-lab.queensu.ca/stat/sas/sasman/sashtml/ods/z0384852.htm
=====================================
Section: The SAS reference:
The general latest
SAS PROC document for reference
http://bama.ua.edu/sasv8/stat/index.htm
This document also
gives ODS dataset names for every Proc STAT step; If you use
this you do not need to use the method given in the above
section.
======================================
Finally my own sweet little codes library: I will be
posting all the niceties learned in the last 15 years of
experience (obviously large application specific codes are not
posted here, but quick sample codes only, but enough number
for novices to use and regular users like me to recall - );
Also, to bring
together all my and my colleagues experience in using SAS and
Statistical Methods, you may visit my Blog:
Statistical Methods and Applications; your
comments are most welcome, and your contributions are most,
most welcome.
======================================
Top 10
questions a statistician should have answers at the fingers
tip.
1. Application
of set theory
2. Principles
of combination and permutation
3. Application
of basic probability problems
4. Application
of Bayesian Theorem
5. Application
of rate of growths and decay problems
6. Application
of three different types of averages, dispersion measures
7. Application
of top 5 probability distributions
8. Application
of linear model principles and top 10 formulae for various
principles in linear models
9. Challenges
and alternative methods for OLS
10. Sampling
problems and design principles
.
|