[prev in list] [next in list] [prev in thread] [next in thread] 

List:       sas-l
Subject:    Re: proc fcmp sql means as additional record on sas dataset
From:       Fried Egg <friedegg2012 () GMAIL ! COM>
Date:       2012-06-28 19:33:11
Message-ID: CANS8eEhyPVhX8eUDg3AZjoWqSeZaSS4YD_H1W6EJSPz129HBTA () mail ! gmail ! com
[Download RAW message or body]

Also,

In your provided example you want to calculate the inverse of a matrix, you
can do this in PROC FCMP as follows:

proc fcmp outlib=work.func.matrix;

 /* wrapper for FCMP Procedure Special Maxtix CALL Routine INV

    INV CALL Routine
    calculates a matrix that is the inverse of the provided input matrix
that should be a square, non-singular matrix.
 */

 subroutine minv ( in[*,*] , out[*,*] );
  outargs out;
  call inv(in,out);
 endsub;

 /* other Matrix CALL Routines here:

http://support.sas.com/documentation/cdl/en/proc/61895/HTML/default/viewer.htm#a003193719.htm
 */

quit;

options cmplib=work.func;

data _null_;
 array mat[2,2] _temporary_ (1.3,1.6,1.1,1.9);
 array inv[2,2] _temporary_;
 call minv(mat,inv);
 put inv[1,1] inv[1,2] /
     inv[2,1] inv[2,2];
run;


On Thu, Jun 28, 2012 at 1:18 PM, Fried Egg <friedegg2012@gmail.com> wrote:

> You should probably invest the time in SAS/IML since it is pretty much
> fully integrated with R, see here for more information:
>
>
> http://support.sas.com/documentation/cdl/en/imlsstat/63545/HTML/default/viewer.htm#statr_toc.htm
>
>
> When using the JAVA object make sure you are using the correct variable
> call type when attempting to retrieve data back to your datastep.
>
>
> PROC FCMP cannot declare data step component objects, so it cannot
> interact with JAVA so your intended interface with this approach would not
> be possible, currently.  The have added support in 9.3 to PROC FCMP to call
> hash methods, I believe, but you still cannot declare new hash objects last
> I heard.
>
> I mean I guess you could have a function that runs a macro that runs a
> data step that runs a java object routine that runs a routine in r that
> returns a value to java that returns a macro variable that returns to the
> function, but god-oh-god why?
>
> Using SAS Integration Techologies you could setup something like SAS <->
> JAVA <-> R where JAVA is really doing the integrating through both
> technologies.
>
> There are many ways to approach any problem, but with the goal being: I
> want to integrate R packages and functions into my SAS programs SAS/IML
> provides a perfectly simple means for doing so.  My 2 cents...
>
>
> On Thu, Jun 28, 2012 at 12:46 PM, Roger DeAngelis <
> rogerjdeangelis@gmail.com> wrote:
>
>> Hi Fried Egg,
>>
>>   Great example. Thanks.
>>
>>   R has a nice low level interface with JAVA. I did try to use the SAS
>> JAVA obj to pass a datastep array to JAVA and then to R and I believe I
>> was
>> able to get the array to R. But I was unable to get anthing back from
>> JAVA/R.
>>
>>  What I tried to do was use SAS datastep->JAVA->R->JAVA->SAS Datastep. But
>> I could not get it to work.
>>
>>  My goal was to wrap all this into a separate FCMP address space
>> subroutine, the java stuff was pretty ugly. All the user would see is:
>>
>> data inverse;
>>  array mat[2,2] (1.3,1.6,1.1,1.9);
>>  array inv[2,2] ;
>>  call inv(mat,inv);
>> run;
>>
>> If R could create a SAS dataset I think I could do the following with
>> FCMP.
>> Note R can read a binary SAS7BDAT.
>>  data inverse;
>>      set mat inv open=defer;
>>      call inverse(mat,inv);
>>      /* should be able to operate on the inverse in here */
>>  run;
>>
>> Years ago - From Richard DeVenezia
>> /* T002320 CALLING JAVA FROM SAS TO CONCATENATE TWO STRINGS */
>>   data _null_;
>>      length s_out $200;
>>      first='ABCDE';
>>      second='FGHIJ';
>>      declare JavaObj j1 ('java/lang/String',first);
>>      declare JavaObj j2 ('java/lang/String',second);
>>      j1.callStringMethod ('concat', j2, s_out);
>>      put s_out=;
>>      j1.delete(); j2.delete();
>>   run;
>>
>>   log should have
>>   S_OUT=ABCDEFGHIJ
>>
>>
>>   I have wasted to much time already on SAS/IML, R is much more powerful.
>>
>
>
[prev in list] [next in list] [prev in thread] [next in thread]