Macro help please

[woo]

Hi Tom -  i am using your code and it is giving me below error message...

____________________________________________________________________________________________________

29            indsname=indsname   ;   where date ne today()-2 ; dsname=indsname ; run

                      _

                      22

                      76

29       ! ;

ERROR 22-322: Syntax error, expecting one of the following: END, KEY, KEYS, NOBS, POINT. 

ERROR 76-322: Syntax error, statement will be ignored.

[Reeza]

Start small, do it manually first, without macros and then convert it to a macro.

Also what version of SAS do you have, indsname isn't available in SAS 9.1 for example, it is in SAS 9.2+.

Change the following so you reference several of your data sets instead of what I have and see if it gives you what you want. Then work from there.

data test ;

  length indsname dsname $256 ;

  set

*list some of our datasets here instead;

dir1.rep1

dir2.rep1

dir3.rep1

  indsname=indsname

  ;

  where date ne today()-2 ;

dsname=indsname ;

run;

[woo]

Hi, i understand this logic at some extent but if below is how i can test code at lower scale then i am receiving same error;

libname rep1 "f:\woo\dir1\test_datasets";

libname rep2 "f:\woo\dir2\test_datasets";

libname rep3 "f:\woo\dir3\test_datasets";

data test ;

  length indsname dsname $256 ;

  set

       rep1.test_check

       rep2.test_check

       rep3.test_check

  indsname=indsname

  ;

  where date ne today()-2 ;

dsname=indsname ;

run;

[Reeza]

Also what version of SAS do you have, indsname isn't available in SAS 9.1 for example, it is in SAS 9.2+.

Post your full log with errors (not just the error) from above code if you're still having errors and are on SAS 9.2+

[woo]

yes i am using 9.1

[Reeza · Accepted solution]

12 year old unsupported software isn't a great idea in general.

Try this:

data test ;

  length indsname dsname $256 ;

  set

       rep1.test_check (in=a)

       rep2.test_check (in=b)

       rep3.test_check (in=c)

  ;

  where date ne today()-2 ;

dsname=a*1+b*2+c*3 ; /*1=rep1, 2=rep2, 3=rep3*/

run;

[woo]

Thank you so much all.

Reeza it worked fine and i was able to convert it into macro. as below. tested - ran fine.

%macro test;

data test;

length indsname dsname $256;

set

%do i=1 %to 10;

"f:\woo\dir&i.\test_datasets\test_check"

%end;

;

  where date ne today()-2 ;

dsname=indsname ;

run;

%mend;

%test

[art297]

Woo: I don't see how that could have provided anything but missing values for both dsname and indsname. Your code didn't follow Reeza's suggested code.  I'd still like to know why the other code versions didn't work.

[woo]

i tested code from my last post and it worked fine. I also did try with different logic like   where date ne today()-3 ; and it worked fine too.

i didn't make any major changes for macro piece of code that i have posted and i still have doubt on (dsname=indsname) (last 4th line of code from what i have posted) so i don't know what line in macro code should replace this dsname=a*1+b*2+c*3 ; from simple data step perspective but i simply used dsname=indsname; and it worked...

[Reeza]

If you're on SAS 9.1 there's no way that works.

[Tom]

If it does what you need then great.  It is hard for me to tell as I still have no idea what you are actually trying to do.

But if you are using 9.1 then it could not have pulled the names of the contributing datasets into the new table since INDSNAME option is not available and you didn't use it anyway.

[woo]

I am using 9.1

I assigned "SAS_MAIN"  -->  right click on it ---> properties   ---> "software" tab and i can see that "SAS Version" is 9.1

[woo]

Tom - this was the question.

i have 10 different directories (like below). under each directory have dataset name "test_check". Our process (other code) producing these "test_check" dataset in 10 directories where we have one variable "date" (type=numeric) which has value like "Sat, Feb 28, 2015" with format "weekdate17." (this value is from today's process - tomorrow it will be"Sun, Mar 01, 2015"). Now basically all the time this dataset, "test_check" has same value for all 10 datasets (i.e two days back: Sat, Feb 28, 2015) but sometime lets say 2 out of 10 datasets has different value, i want to find out in which directory "date" variable has different value than lets say Sat, Feb 28, 2015 (if i run code today).

__________________________________

f:\woo\dir1\test_dataset\test_check

f:\woo\dir2\test_dataset\test_check

.

.

.

f:\woo\dir10\test_dataset\test_check

[Tom]

If you just have 10 directories to check then forget macro code and just write the program. Use your editor to replicate the lines that are similar.

To find out if the dates do not match each other you don't need to know if they match today's date.

data check ;

    merge

     "f:\woo\dir1\test_dataset\test_check" (rename=(date=date1))

     "f:\woo\dir2\test_dataset\test_check" (rename=(date=date2))

....

     "f:\woo\dir10\test_dataset\test_check" (rename=(date=date10))

    ;

    if N(of date1-date10) < 10 or min(of date1-date10) ne max(of date1-date10);

    put (date1-date10) (=/);

run;

[woo]

Tom it worked great but can you please brief me about

why we put this sign?(=/);