The easiest way is to allocate a matrix of missing values and then fill in the nonmissing values by using subscripts. Suppose your data look like this: N = 4; /** number of rows after including missing values **/ MissingIdx = {2, 3}; /** indices of rows that contain missing **/ x = {1,2}; /** data values **/ The hard part is that you need to know the indices of the NONmissing observations, which you can find by using the SETDIF function. The function is explained at http://support.sas.com/documentation/cdl/en/imlug/59656/HTML/default/langref_sect241.htm NMidx = setdif(1:N, MissingIdx); /** find the indices for nonmissing values **/ newX = repeat(., N); /** create answer vector with all missings **/ newX[NMidx] = x; /** assign nonmissing values in the specified locations **/ print newX; ... View more

I can tell that you have a MATLAB background. Welcome to IML! In IML, use a space to separate columns and use a comma to indicate a new row. So your data might look like this: A = {1 2 3, 4 5 6, 7 8 9}; x = {1, ., ., 4, ., 6, .}; The function you want is the LOC function, which can find the indices of x that are missing (or nonmissing): NMidx = loc(x ^= .); /** non-missing index **/ v = x[NMidx]; /** subset to form nonmissing values of x **/ y = repeat(., nrow(x)); /** fill answer vector with missing values **/ y[NMidx] = A*v; /** put answer in correct locations **/ print y; ... View more

You can use the "alternative commment" (asterisk ... semicolon) at the end of lines, like this: proc iml; A = {1 2, 3 4}; * matrix assignment statment; x = {5, 6}; * vector assignment statment; y = A*x; * matrix multiplication; print y; As Hutch points out, they are also valid on separate lines. The matrix operator is in the middle of a statement, so the IML parser can distinguish that it is not a comment. It sounds like the color-coding shown in your editor is getting confused by the alternate comment syntax. Just ignore the color-coding, or use traditional slash-star comments (/* ... */). ... View more

The variable names in the ODS tables are obviously being hard-coded as VAR1-VAR16, and I don't think you can change that portion of the output. The problem, of course, is that the LMS routine takes a matrix of explanatory variable (your 'a' matrix), so it actually doesn't know the names of the data set variables from which the matrix was created. It would be a good idea to extend the LMS routine to provide the functionality you want. You can, however, use ODS output to save the tables to a data set and then use the DATA step or PROC IML to rename the variables. If you are interested in that, here is some sample code: %let input = MPG_Highway; %let model1 = EngineSize Horsepower Weight Length; proc iml ; use Sashelp.cars; /* you'll have this if you have SAS/GRAPH */ read all var {&input &model1} into model1; close Sashelp.cars; a = model1[,2:ncol(model1)]; b = model1[,1]; optn = j(8,1,.); optn[2]= 1; optn[3]= 1; optn[8]= 1; *ods trace on / listing; ods output EstCoeff=ParamEst; /* write one or more tables to data sets */ call lms(sc,coef,wgt,optn,b,a); ods output close; /* read data with names 'VAR1':'VARn' and 'Intercep' */ use ParamEst; read all var _NUM_ into x[colname=oldVarNames]; close ParamEst; /* rewrite data with new names */ varNames = {&model1} || "Intercep"; create ParamEst from x[colname=varNames]; append from x; close ParamEst; quit; proc print data=ParamEst; run; ... View more

1) Yes, you can call SAS procedures from SAS/IML Studio. I suggest you download the PDF for SAS/IML Studio for SAS/STAT Users at http://support.sas.com/documentation/cdl/en/imlsstat/63827/PDF/default/imlsstat.pdf Chapters 4-7 deal with common tasks such as calling SAS procedures and importing the data into SAS/IML Studio. 2) The R interface was just announced last year, but there is a SAS Global Forum paper that presents examples that go beyond the documentation: http://support.sas.com/rnd/app/papers/papers_iml.html The 2010 paper has examples of calling both SAS and R. The 2008 paper features calling SAS. There is also a new book on IML and IML Studio coming out, but, alas, it is not yet available: http://support.sas.com/publishing/authors/wicklin.html 3) At this point, SAS/IML Studio is the only SAS product that has an interface to R. (The next release of PROC IML will also feature an interface to R.) Your "plan" on how to use IML Studio and call R when needed sounds right. That is essentially the way that I use the product for data analysis and the development of statisical algorithms. ... View more

As explained in a previous thread, some Base SAS functions such as RANPERM expect a list of arguments instead of a vector of arguments. See the thread http://support.sas.com/forums/thread.jspa?threadID=7763&tstart=0 If you know that you have three objects to permute, you can use RANPERM like this: proc iml; x1=1; x2=2; x3=3; seed=12; /* or use 0 for random seed */ call ranperm(seed,x1, x2, x3); print x1, x2, x3; However, there is an easy way to create a permutation of a vector of an arbitrary size: generate random uniform numbers and then compute the RANK of those numbers to use as an index vector. For example: x = {1,2,3}; call randseed(12); u = j(nrow(x),1); /* allocate vector */ do i = 1 to 5; call randgen(u, "uniform"); /* fill u with random uniform */ perm = x[ rank(u) ]; /* use RANK(u) to permute data */ print i perm; end; ... View more

You need to generate a permutation of the numbers 1:N (N=number of obs) and use the permutation as indices to form a new vector. There are several ways to do this. The simplest is to generate random uniform numbers and then use the RANK function to assign integers that correspond to the order of the random numbers: N = 10; /* number of observations in data */ u = j(N, 1); /* allocate Nx1 vector */ call randgen(u, "Uniform"); /* fill u with random uniform */ p = rank(u); /* p is random permutation of 1:N */ print u p; newY = oldY ; /* new vector = permutation of data */ ... View more

I think the fundamental problem is that the error occurs in the EM module, but you are trying to jump to a statement outside of the module. that's why the ST label is not recognized. What I like to do is to have the module that might fail return a "return code" that can be handled by the calling routine. In the code below, I use a DO/WHILE loop to jump out of the iterative loop when the EM module fails. I then call the INIT routine and continue to the next set of starting values. proc iml; start init; x=1; finish; start EM(args); onError={ "if ErrorChecking then do;" "rc = 1;" "call push(onerror);" "resume;" "end;" }; call push(onError); rc = 0; /* return 0 if no error */ ErrorChecking = 1; /* turn on error checking */ t = log(0.75-uniform(54321));/* error 25% of the time */ ErrorChecking = 0; /* turn off error checking */ return (rc); /* return whether error occurred */ finish; myargs = {1,2,3}; run init; do s=1 to 3; err = 0; /* no error */ do e=1 to 5 while(err=0); /* exit loop on error */ err = EM(myargs); /* perform 1 iteration*/ print s e err; end; if err then do; /* error: reinitialize */ print "*** Error occurred...reinitialize ***"; run init; /* regenerate starting values*/ end; end; ... View more

Rats: The blog software seems to "eat" certain character sequences. Very annoying. The statement inside the DO loop should say: args = concat(args, ",", strip(cc [ i ])); /* LEFT_BRACKET i RIGHT_BRACKET */ ... View more

As model_coder mentions, there is a typo in the program. The statement inside the DO loop should say: args = concat(args, ",", strip(cc)); ... View more

Yes, R_HOME should point to the installation directory, not to the `bin` subdirectory. The main purpose of R_HOME is to enable SAS/IML Studio to work with a particular version of R when you have multiple versions of R installed. When you install R from Windows binaries (for example, by launching `R-2.10.1-win32.exe`) it automatically sets R_HOME (or maybe there is a checkbox in the installer), so most folks don't need to fiddle with R_HOME at all. I'm curious. Was R_HOME not set properly after you installed R? Did you install from Windows binaries or did you build R from source code? ... View more

I assume that you are able to run R from Start-->All Programs-->R? If so, I think you should contact SAS Technical Support and send them the following information: 1) The version of R you are using 2) The version of Windows you are using 3) Are you using a 32-bit or 64-bit version of Windows? 4) Are you using a 32-bit or 64-bit version of R? ... View more

Of course. Sorry. The variances are in v; the covariances are in low. ... View more