Monday, April 27. 2009Creating a SAS format from a data set
Recoding values is one of the most common data prep tasks that folks need to do before they can analyze and report on data. In SAS, the most elegant way to handle this is by applying a SAS format. A SAS format allows you to "bucket" a bunch of raw data values into more managable categories. For example, a professor can map numerical test scores into letter grades ("A", "B", down to "F"), or even further into simply "pass" or "fail".
You can create a SAS format from scratch, writing a SAS program to include name-value mappings for recoding values, or ranges of values to map to a particular formatted label. But often you already have this information in a data set or spreadsheet, and it would be easier if you could just read in those values and make a SAS format from those recoding rules. There is plenty of documentation and also conference papers to describe how to use the FORMAT procedure to get this done with SAS programming. The SAS Education team considers it an important topic and they regularly teach it in their beginning SAS courses. The trainers who cover SAS Enterprise Guide have been asking, on your behalf, for an easy way to do this without having to dive into SAS programming. The SAS educators commissioned this task, and we've built it. It runs as an add-in to SAS Enterprise Guide 4.2. The remainder of this post describes the details of the task, where you can download it, and how you can use it.
Where to downloadThe task is a custom task built for use in SAS Enterprise Guide 4.2. (It does not work in SAS Enterprise Guide 4.1 - sorry!) I've posted a version of this on the SAS support site, here. You can download the ZIP file and save it to your machine. The instructions for installing the task for use are in the README.txt file within the ZIP file.
Scenarios supportedFor example, imagine you want to categorize US states into their broader regions, like so:
Or perhaps recode ranges of values into discrete labels, like this:
Screen shotsThis is a picture of the task when used to create a "look up" format, where discrete values are mapped to formatted labels. In this example, the data is mapping product IDs to product names.
This is a picture of the task when used to create a range-based based format, where you select lower and upper values in a range to map to a formatted label.
Trackbacks
Trackback specific URI for this entry
No Trackbacks
Comments
Display comments as
(Linear | Threaded)
Cool. I built a similar app a few years back for DSUG. It's nuance was that it could take any datasource (OleDb, Excel, etc.) and automatically generate the formats for you. Used drag-and-drop (Infragistics datagrid).
My point was to illustrate that SAS code could be generated outside of macros. May be a feature to add in at a later point. Alan I frequently have layouts of text files or lists of variable values I need formatted in an Excel spreadsheet. I find that simply using the Excel CONCATENATE function that allows me to to write out the SAS code in Excel and then drag that formula down throughout the entire workbook is a very fast way to generate the SAS code I need. I just click on the Excel column in which I generated the SAS code and paste it into SAS.
This works for much more than format statements. Inelegant perhaps -- but accurate and very fast. You are my EG super-hero. If I had a nickle for every time I had an EG student ask me for this, I'd be a rich man!
Now I just have to wait until my students upgrade to 4.2  Andy Credit where credit is due: the task was developed by my colleague in New Zealand:
http://blogs.sas.com/sascom/index.php?/archives/295-Five-questions-with-David-McNamara.html Chris |
ABOUT CHRIS Chris Hemedinger is a senior software manager in SAS R&D. He's part of the team that builds SAS Enterprise Guide and SAS Add-In for Microsoft Office. He's also co-author of the popular SAS for Dummies book and moderator of the SAS Enterprise Guide discussion forum.
QuicksearchSyndicate This BlogOther SAS BlogsBI and Chicken Pot PieConversations and Connections Data Flux Community of Experts Closing the Intelligence Gap In Other Words In the Final Analysis JMP Blog Key Happenings at support.sas.com The New Lending System The Text Frontier SAS Global Forum Blog sascom Voices Blog A Shot in the Arm All SAS Blogs Recent postsUsing the X and SYSTASK commands from SAS Enterprise Guide
19Nov2009 Learn about ODS Statistical Graphics 06Nov2009 Hooo-ray for Hollywood 03Nov2009 Where to find ODS Statistical Graphics in SAS Enterprise Guide 23Oct2009 Welcome sasCommunity visitors! 21Oct2009 Do me a favor: stop doing me favors 20Oct2009 You can use it for evil, too 16Oct2009 The CIO perspective from MWSUG 13Oct2009 SAS 9.2 and SAS Enterprise Guide 4.2 on Windows 7 01Oct2009 SAS 9.2 support for Windows Server 2008 30Sep2009 BOOKS & VIDEOS SAS for Dummies -- If you want to learn to be productive using the latest SAS tools and technologies in your daily work, but don't necessarily want to learn SAS programming, then this book was written for you!  Efficient Data Access with SAS Enterprise Guide (YouTube) -- Shows the best way to access data in a DBMS from SAS Enterprise Guide, using M&Ms as a prop.  Exciting combinations: SAS and Microsoft Office (YouTube) -- A metaphor for the exciting features you get when you add the SAS Add-In for Microsoft Office to your desktop.
The blog content appearing on this site does not necessarily represent the opinions of SAS. Your use of this blog is governed by the Terms of Use.
|
