Erwan,

Country is indeed undefined when you call the function modNC. I think you are mixing your variable names and literals.

Try something like

NCou=modNC({country, n});

for the function call. And then modify the read statement to

read all var arg1 into NC where (_type_='100');

In your version, arg1 is part of the literal and so the read statement would fail when it checked to see if there was a variable called arg1 in the test1 dataset.

By the way, if country is text and n is numeric then you will have more problems since an IML matrix is either all character or all numeric.

Ian.

-----Original Message----- From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Erwan LE DU Sent: 01 August 2008 09:07 To: SAS-L@LISTSERV.UGA.EDU Subject: Debugging problem in proc IML

Hi all,

I have an issue with the following IML procedure. I have a dataset test1 which is called from modMain by module modData. I then call the modNC (module with argument) which is called from modMain using the statement :

NCou=modNC(Country);

The issue is there : I get the following error message (error) Matrix has not been set to a value.

However Country is a valid variable in the dataset test1.

I tried nesting modData and modNC within modMain without success.

Can someone help me on this?

Regards,

Ker

*

proc **iml*;

reset noprint;

reset log;

reset flow;

reset details;

start modData;

mydata=concat('use test',char(*1*,*1*),';');

call execute(mydata);

finish modData;

start modNC(arg1);

read all var {arg1 N} into NC where (_type_='100');

return(NC);

finish modNC;

start modMain;

*run* modData;

NCou=modNC(Country);

print NCou;

finish modMain;*

run *modMain;

quit;