With such a general question, I can only offer general ideas. -- generate the complete data for the simulation study. -- use a data step and a random number generator (ranuni, say) to identify data points to make missing. This has the disadvantage of only generating missing data randomly. That means that analyses will work just like you have complete data with an effectively lower sample size. If you want to make the missing data dependent on other variables, you can just nest the missing generation in IF statements. That dependency is the other extreme of missing data analysis. You may wish to read the literature on analyzing missing data, as that can help in the strategy for creating missing data to meet your analytic goals. There are probably some references in the PROC MI chapter of the SAS/Stat manual. ... View more

You may be able to find it discounted as a remainder, as SAS has discontinued the product. Macro programming is part of base SAS, so you will have it. SAS is a licensed product. When the license expires, it stops working. That is standard for all SAS products. Try Amazon or other online retailers, they often have used books. You can learn a lot of SAS with one or two books (search the forum for recommendations). The SAS reference manuals are available for free at support.sas.com. Doc Muhlbaier Duke ... View more

Welcome to the forum. You need to frame your questions with more context to get sensible answers. For instance, without further context, an appropriate answer to question 4 is "the same way we use it on the other SAS platforms." It's true, but I don't think that it is useful. Doc Muhlbaier Duke ... View more

Two things: On SAS 9.1.3, only two CPUs are supported for multi-threading unless you are running windows server edition and SAS for windows server. http://support.sas.com/kb/12/568.html Not all procedures support multi-threading. I've read a list of which do and how many processors max, but couldn't find it now. Maybe someone else will have it. You may find that SAS 9.2 has updated the multiprocessor support; check with tech support. Doc ... View more

PROC SETINIT will tell you what software you have licensed. The best way to determine what is available on a Windows machine is to run filedetvX (x is 8 or 9), see http://support.sas.com/kb/7/640.html You can run that from within EG if you are setup to have access to the "X" command (that won't work for SAS on a remote server). Doc Muhlbaier Duke ... View more

EGuide, by default, uses the OPTION VALIDVARNAME=ANY. If you use that variable in a procedure that does surface the code, you will see that the variable name is referenced as "2008"n . You also need to reference non-standard names that way. See the SAS/Base manual for the section on reverencing variable names. ... View more

SAS uses ODS to deliver the output to EGuide. EGuide builds some ODS statements into the code it sends to the SAS server. SAS has an entire book on ODS in the documentation at support.sas.com. It can get complex enough that there is also a SAS Discussion Forum devoted to it. I'm no ODS expert (far from it), so I'd just try cut-n-paste on the code that EGuide generates to see if that will work reasonably before delving into the documentation too deeply. ... View more

Be careful with really long numbers. SAS stores all numbers in floating point and can only represent integers to certain level of precision. See the companion for your OS to determine the maximum length of an integer that can be represented for a particular number of bytes of storage. ... View more

"Characterize data" does allow you to set a limit for the number of distinct values that it shows (the default is 30). It displays the **first** k distinct values, which can be misleading if there is some inherent order in the data that you don't know about. A separate limitation to "Characterize data" is that it only does the frequencies on character variables; it does mean/medians/etc. on numerics. Maybe EGuide 4.2 will be smarter. ... View more

My guess is that SAS is not using the old data, but rather there is no output from the procedure and the report does not get replaced. There are two ways to check for this. 1) look at the time stamp in the footer. It shows when the table was created (the one in the header just shows today's date, so it is not of much use). [You may need to modify the titles tab to put the footnote back in.] 2) look at the task log. If it has errors, the ODS CLOSE statement is not run and the report is not replaced. I'm not sure how to solve this, but here is a guess. Add some ODS code to the beginning of the task to open the output destination and then close it. Then open it again in the regular way. I've not tried it, but I suspect that it will create a null report (perhaps just titles) that is replaced with a successful report if the task completes. ... View more

Agree with Andre. UNIVARIATE is such an old procedure that it hasn't kept up with the flexibility of the other graphics in SAS. It doesn't even handle ActiveX for ODS RTF output properly (9.1.3 is smart enough to change the graphic call to something that UNIVARIATE did understand; 9.2 lost that work-around). ... View more

The problem is elsewhere in your program. MEANS uses all the data it gets. If the variable in the VAR statement has a missing value, then it does not count toward the mean or other statistics (except N and NMISS). If you are losing duplicates somewhere, it is often in a FIRST. or LAST. construct in the DATA step or in the DISTINCT option of SQL. Without more information from you (code, log, lst), I'm at a loss to help further. ... View more

There are several places in SAS that involve multivariate normality. What procedure are you using? Reading the references under the procedure "details" often tells how to interpret the statistics. ... View more

Enterprise Guide uses SAS in a client/server configuration, so somewhere you have the complete SAS. SAS has several products that include the word "Enterprise" as an adjective. In Enterprise Guide, you can use the "Characterize data" task to get a quick view of a dataset. It won't do all you want, but it will give you a high level overview of each file. Doc Muhlbaier Duke ... View more

Doing what you describe takes some sophisticated programming using the ANNOTATE facility in SAS. It also makes for a distorted visual image (see Cleveland's graphics book). An alternative that is much easier and not distorted is to concatenate the salary to the name, like "venkat, 5000", and then do a horizontal bar chart with the DISCRETE option (I'm not sure that it will display 500 bars). ... View more