by Richard A. DeVeneziaSend Feedback
These code samples have been culled from various work experiences, programming experiments and snippets posted to newsgroup comp.soft-sys.sas (SAS-L)
These sources are presented as-is. Use at your own risk.
|report hbar, sudoku, spinward gradient, spinward animgif, jumble, dairy-herd, 200 plots, apache log, amicable numbers, hash-8, joy, hash-7, jDSGI,|
|🖹 dir.c||C program to output filesize, file modification date (m/d/Y T), and filename. Useful for processing groups of input files.|
|🖹 hex.c||Hex dump a file for detailed examination.|
|🖹 hexcat.c||Similar to hex, but honors newlines found in file.|
|🖹 mac2ux.c||Pipe fitting to convert files with \r line termination to \n line termination.|
|🖹 mdate.c||Return modification date of a file.|
|🖹 windows-ipaddress||Windows - Determine TCP/IP internet address
Posted to SAS-L Nov 03, 1999.
Adapted from Randy Birch code at VBnet
|🖹 windows-no-x-icon||Windows - Disable Close menu item (and X icon) in the system menu of a SAS session
|🖹 splash-countdown||Generate jpg images and display them in a borderless window as a series of splash images. Requires windows program Splash.exe.|
|🖹 uga.vbs.txt||Small visual basic program to open the listserv.uga.edu web interface to SAS-L, showing newest posts first. This code could easily be adapted to a bookmarklet.|
|🖹 constant||Demonstrates lack of optimization by DATA Step compiler. Advice: Don't use
|🖹 date-formats||Lists the output of every SAS date related format. The log.|
|🖹 read-html-data||sas program to run perl script so it can read data extracted from web page|
|🖹 uuencode||uuencode implemented in a SAS DATA Step|
|🖹 read-web-log||Use SAS to read an Apache web server log file in common format|
|🖹 beers-n-bars||Which bars sells the beers a drinker likes ? Or Which sets contain which subsets. SQL and bitcoding.|
|🖹 bit-count||Count the number of bits set to 1 in an integer value|
|🖹 bit-reverse||Reverse the order of the bits of an integer value|
|🖹 cartesian-join||From a set of N points, keep only the pairs of points which are closer than some threshold. A Data step solution and an exquisite SQL solution by Howard Schreier.|
|🖹 dairy-herd||Determine which calves were exposed to shedding dams|
|🖹 hash-1||DATA Step hashes performing a frequency analysis|
|🖹 hash-2||Rediscover the paths traversed by wandering subjects and report on them|
|🖹 hash-3||Some pitfalls you might encounter. Read access violation and missings turned into zeros.|
|🖹 hash-4||Transpose and concatenate satellite data while key merging disordered sets.|
|🖹 hash-5||Sum amounts in a transactional trail.|
|🖹 hash-6||Split a table into subset tables based on variable values aka Split one data set into several data sets named according to a group
Material posted in 24MAY2004 SAS-L thread "Hash of hashes and dynamic output"
|🖹 hash-7||Hash of hashes idea (see hash-6) expanded into a useable multi-level-hash (MLH) construct by using macros. See
multi-level-hash for the macros. An MLH in essence provides for arbitray navigation of the class
levels of a hierarchy within the providence of a DATA Step.
Important: Any PROC Step that provides a CLASS or BY statement allows you to process data that is hierarchically segregated into groups. Use Procs instead of MLH in most every case.
Additionally, in many DATA Step situations that involve hierarchical data and hashes, consider composite keys (one hash with >1 key variable) instead of MLH. (There are probably some theorems in computer science regarding the equivalences of one multi-keyed hash to N hashes that partition the key set.)
I have yet to find a reasonable sample problem that demonstrates MLH as the principally simplest solution technique. If one is found it will probably involve atypical path following algorithms or recursive searches. MLH is a technique waiting for a problem.
|🖹 hash-8||In a set of dual keyed things find the groups of things chain related by key1 or key2, aka
A response to 17AUG2004 SAS-L thread "How to group people by their first name OR last name"
|🖹 how-much-hotter-it-got||A pot can contain 20 components. The temperature is known for every combination of components. How much hotter does the pot get when one new component is added to the pot?|
|🖹 increment-every-other||Increment a counter every other observation.|
|🖹 index-within-group||Seven techniques for calculating a within by group sequence number (What row in this group is this observation ?)|
|🖹 _infile_ buffer||Input data directly from a character expression by managing the _infile_ buffer. (Sort of like C's
|🖹 jumble||Solve the popular newspaper word puzzle|
|🖹 least-likes||A collection of like clauses likely overlaps. What are the fewest like clauses that would match the same things as all the like clauses.|
|🖹 mapped-lookup||A DATA Step processing an atypical data structure requiring the assignment
|🖹 polygon-hunt||An unordered data set of line segments having unique vertices are known to be pieces of polygons. Enumerate the vertices of each polygon and indicate if the order is clockwise or counter-clockwise.|
|🖹 random-schema||Create a SAS library containing a complicated schema involving dozens of tables and dozens of referential relationships.|
|🖹 rolling-back-sums||At time
|🖹 rxparse-in-macro||How to use rxparse, rxchange and rxfree in macro|
|🖹 split-into-words||Split a long character value into several shorter character variables (splits) while ensuring the splits contain only whole words|
|🖹 superset||Find the fewest mutually exclusive supersets S* such that each S* contains all sets S sharing at least one member. Demonstrates how to
program a recursive algorithm using GOTO and ARRAY to emulate functional calls with stacks.
|ODS - Output Delivery System|
|🖹 pdf-bullets||Characters in ITC Zapf Dingbats font make excellent pdf bullets. Generate a sample showing each.|
|🖹 pdf-bullets||Place bulleted text in a pdf document.|
|🖹 excel-multisheet||Use windows shell script to create a new Excel workbook containing several worksheets, each populated from an html file. Each html file is generated using ODS and contains Proc TABULATE results for one by-group.|
|🖹 report hbar||Generate hbar charts per by group using Proc REPORT and ODS. Sample output.|
|🖹 tabulate||Proc TABULATE - Use the page dimension of tabulate to place a title above each table created in a single Proc step. Sample output.|
|🖹 tabulate-with-graph||Generate a two page report in pdf format. Contains a tabulation and vertical bar charts.|
|🖹 annontated-chart||A stacked vbar chart annotated to display counts of another subgroup.
|🖹 bezier-curves||Bezier curves inside a circle. (DSGI)
|🖹 cowboy-hats||Use GREPLAY to print 200 plots, 2 plots per page.
Sample uses 200 cowboy hat plots.
|🖹 font-dump||View all characters of any or all SAS fonts. Windows platform can also examine true type fonts. Notes: SAS font samples can also be found in the OnlineDoc. Special symbols that can be indicated in a SYMBOL statement can also be found in the SPECIAL font.|
|🖹 font-zodiac||Zodiac and planetary symbols are found in the SAS SPECIAL font.
|🖹 G3D-annotated||Surface function plot annotated with data points. Change in view point animated using gifanim graphics device. Sample output.|
|🖹 gchart||A stacked vertical bar chart with groups.|
|🖹 gfont-quadrants||SAS/Graph custom font created with Proc GFONT. The sixteen possibilities of a circle with it's quadrants filled and unfilled. This font
uses circular arcs, which in my estimation contains too few sampling points. A thumbs down on circular arcs.
|🖹 gfont-moon-phase||SAS/Graph custom font created with Proc GFONT. Shows N phases of the moon. You choose N. This font does it's own arc computations and at
small sizes looks better than if SAS circular arcs were used.
Java DATA Step Graphics Interface
A combination of Java classes and SAS macros that let you can create almost any type of static 2D visualization as an image file (PNG and JPG are built into JRE). The impetus at the time of this development was the lack of antialiased line drawing in SAS/Graph.
Additional graphic formats are supported by adding ImageIO jars to your JRE lib/ext folder. See
Java Advanced Imaging and download
I/O Tools. Windows users must set an undocumented environment variable
for their SAS sessions. Use command line option
|🖹 l1-traveling-salesman||A traveling salesman, cities on a grid and shortest path to next city. Cities can be made to exist on a spiral :).
|🖹 month-o-charts||A month's worth of bar charts in a html page|
|🖹 schema-ball||Graph the foreign key relationships of tables in a library. (DSGI)
Inspired by Martin Krzywinski's schemaball.
Run the random-schema program to fill a library with randomly related tables.
|🖹 state-chain||An NPR puzzle: The last two letters of ALASKA are the first two letters of KANSAS. Can you find a chain of five state names that overlap like this, in which the last two letters of each are the first two letters of the next. See the answer.|
|🖹 vbar-by-group||Ensure all midpoints are same patterned even when some by groups do not contain all midpoints found across all the by groups. Sample output.|
|🖹 amicable-numbers||A pair of numbers, the factors of each when summed is the other.|
|🖹 descent||An n-sided regular polygon rotating a fixed amount and shrinking to be circumscribed by the previous iteration. Iterations are colored according to a linearly interpolated rgb gradient. (DSGI)|
|🖹 joy-to-the-world||Generate wave forms of pure tone notes and play them back through the audio device in the form of the song "Joy to the World". Tested on Windows 2000.|
|🖹 sudoku||Solve Sudoku puzzles using a path walker. Simple front end for tracing through each step.|
data fakedata; x = 0; do while (x<100); y = ranuni(0); output; * increment by one or two; x + 1 + ranuni(0)<0.25; end; run;A. What is the most number of rows table fakedata could have ?
B. What is the least number of rows table fakedata could have ?
These sources are presented as-is. Use at your own risk.This page was last updated 30 May 2007.