DET is not cheap. If the point of your computation is to solve a linear system, why not use SOLVE or INV? You're planning to abort if the matrix is nonsingular, anyway, so just try to solve the system. For a difference between SOLVE and INV, see p. 50-51 of Statistical Programming with SAS/IML Software, which is available as a free download at http://support.sas.com/publishing/authors/wicklin.html (Also, Ch 15 discusses performance considerations between the two functions.) Your code, as written, probably won't detect most matrices that are numerically singluar. For example, if DET returns 1E-100, your code won't set the error condition. Rick Wicklin SAS/IML Blog: http://blogs.sas.com/iml ... View more

1) Yes, the overflow means that the determinant is so large (in magnitude) that it cannot be stored as a double precision number. This happens often for large matrices. Your matrix dimensions are 145x145. If you take a matrix that large and fill it with uniform random numbers between 0 and 10, the determinant might be 1E200. For example: d = det(10*uniform( j(145) )); 2) See Chap 18 of the IML doc, and especially http://support.sas.com/documentation/cdl/en/imlug/63541/HTML/default/viewer.htm#imlug_genstmts_sect010.htm ... View more

Well, if that's really what you want, you can wrap a little module around the loop like this:  start foo; do k=1 to 50;    z=1/( 1+EXP(-W*(X`)) ); /*I want to see this result at every k value*/    pause; end; finish; run foo; Alternatively, just print all the values and then scroll up in the output window.   2) You might want the SHOW command: http://blogs.sas.com/iml/index.php?/archives/142-Listing-SASIML-Variables.html   3) In SAS/IML Studio (which is good for beginners), you can use the pause in "open code" (=not in a module). See http://blogs.sas.com/iml/index.php?/archives/123-How-to-Learn-SASIML-Five-Resources-for-the-Beginner.html ... View more

You say you want to use GCHART, but if you decide to use the newer statistical graphics (SG) procedures, then see http://blogs.sas.com/iml/index.php?/archives/154-Overlaying-Two-Histograms-in-SAS.html ... View more

You can now use the SAS/IML 9.3 doc: http://support.sas.com/documentation/cdl/en/imlug/64248/HTML/default/viewer.htm#imlug_langref_sect022.htm ... View more

If you have SAS/IML 9.22 software or later (after Nov, 2010), you can use the SUBMIT/ENDSUBMIT statements to call SAS procedures and DATA steps from PROC IML without losing the state of your IML session: submit; DATA SAMPLE_SET_A;    ... DATA SAMPLE_SET_B;    ... RUN; endsubmit; See http://support.sas.com/documentation/cdl/en/imlug/63541/HTML/default/viewer.htm#procs_toc.htm ... View more

I'll add that I've typed half of a response, gotten interrupted, and by the time I finish the response and hit "Add Reply" I've been logged out! My response is now g-o-n-e. Nothing burns worse then spending a sizeable amount of time crafting a thoughtful reply, only to have it vanish into the ether. ... View more

This is a root-finding problem. Define f(lambda) = target - cdf("poisson", n, lambda); and solve for the zeros of f. This is always possible (and has a unique solution) because the CDF is an increasing function of lambda. (Of course, 0<target<1). To solve the problem, find a root-finding algorithm such as the bisection method implemented in SAS. (I believe that the %inv macro at http://ftp.sas.com/samples/A57496 gives you what you need; Although it is written as a macro, you ought to be able to decipher it.) If you have SAS/IML, see http://support.sas.com/documentation/cdl/en/imlug/63541/HTML/default/viewer.htm#imlug_genstatexpls_sect004.htm ... View more

I agree with Murphy, but use the latest 9.22 documentation: http://support.sas.com/documentation/cdl/en/imlug/63541/HTML/default/viewer.htm#titlepage.htm Also, see the blog post: How to Learn SAS/IML: Five Resources for the Beginner ... View more

Depends if the function is one-variable or several variables. Also if polynomial or general nonlinear equation. 1) For one variable and polynomial, use POLYROOT. This finds real and complex roots. 2) For finding the real roots of general functions, see Example 9.2 of the SAS/IML doc: http://support.sas.com/documentation/cdl/en/imlug/63541/HTML/default/viewer.htm#imlug_genstatexpls_sect004.htm The Newton-Raphson method (2) will find one root for each initial condition, so if you are trying to find multiple roots, you might want to vary the initial condition. ... View more

I don't understand what the SQL code is trying to do. Is this some kind of resampling? Please describe in words what you are trying to accomplish, or provide the macro function that you are calling so that we can run it ourselves. If this is a resampling problem, I would do it all in SAS/IML without bothering with SQL. For example, I'd read in the ID variable, use the UNIQUE function to get the numbers {1,2,3}, and then use RANDGEN to generate random numbers in the range 1:3. I can't tell yet how you are using the X variable. ... View more

Yes, I know that my code doesn't have this problem, but my code DOES have the same form as your code: the WHERE condition is changing in a loop. I can't reproduce your problem, so I was hoping that you might be able to get it to happen with a data set that we both have access to. ... View more