topic Re: Capture error record in macro variable from a dataset in SAS Programming

Capture error record in macro variable from a dataset

David_Billa — Thu, 03 Oct 2019 12:28:57 GMT

I've written a macro program as follows. First program to create a dataset named 'protocol' and the second macro program is to check for the count in 'dups' dataset.

If the count is greater than 0 in 'dups' dataset then I would like to call the first macro program with some values. In the list of Parameter macro variables there is a macro variable called 'flt_rec_txt'. In this variable (don't worrry about length) I want to feed the values by populating all the observations from 'dups' dataset seperated by space. However I'm not certain how to achieve it in my apporach and I got error as below. Also I was asked not to change the apporach in order to resolve the error.

%macro IFR_WRITE_ERROR(
			error_code =
			, error_text =
			, flt_id =
			,flt_rec_txt =		
			);

	data protocol;
		ERROR_CODE = &error_code.;
		TEXT = "&error_text.";
		flt_id=&flt_id.;
		flt_rec_txt="&flt_rec_txt.";
	run;

%mend IFR_WRITE_ERROR;

%macro IFR_CHECK_KEYS;

	proc sql noprint;
		select count(*)
			into :dup_columns
				from dups
		;
	quit;

	%if &dup_columns. > 0 %then
		%do;
			%IFR_WRITE_ERROR(
				error_code = 100
				, error_text = Validation key - unique key violated
				,flt_id = 128
				flt_rec_txt=dups
				);
			%let processing_code = 1;
		%end;
%end;
%mend IFR_CHECK_KEYS;

%IFR_CHECK_KEYS;

Error message in the log:

MPRINT(IFR_WRITE_ERROR):   TEXT = "Validation key - unique key violated";
SYMBOLGEN:  Macro variable FLT_ID resolves to 128     flt_rec_txt=dups
NOTE: Line generated by the macro variable "FLT_ID".
2433       128     flt_rec_txt=dups
                   ___________ ____________
                   22          557
ERROR: DATA STEP Component Object failure.  Aborted during the COMPILATION phase.
ERROR 22-322: Syntax error, expecting one of the following: !, !!, &, *, **, +, -, /, ;, <, <=, <>, =, >, ><, >=, AND, EQ, GE, GT, 
              IN, LE, LT, MAX, MIN, NE, NG, NL, NOTIN, OR, ^=, |, ||, ~=.

If the Dups dataset has the dataset as below,

FLT_LAYR_NM	FYP_NM	_TEAM_NM
DIS	Consistencycheck	Business
DID	Rangecheck	Client

Then the desried result of the above program should be,

error_code	error_text	flt_id	flt_rec_txt
100		128	DIS Consistencycheck Business DID Rangecheck Client

Re: Capture error record in macro variable from a dataset

novinosrin — Thu, 03 Oct 2019 12:44:33 GMT

Hi @David_Billa A quick look tells me you are missing a comma(,) here between the keyword parameters ,flt_id = 128
flt_rec_txt=dups

So will this correction help?

%IFR_WRITE_ERROR(
error_code = 100
, error_text = Validation key - unique key violated
,flt_id = 128

,flt_rec_txt=dups
);

Re: Capture error record in macro variable from a dataset

David_Billa — Thu, 03 Oct 2019 13:00:32 GMT

No, it's a typo error and it didn't helped me either

Re: Capture error record in macro variable from a dataset

Tom — Thu, 03 Oct 2019 13:06:12 GMT

So post the new error, as the old error was clearly caused by having wrong value passed to the parameter because of the missing comma.

Re: Capture error record in macro variable from a dataset

Tom — Thu, 03 Oct 2019 13:08:36 GMT

It looks like you want concatenate multiple rows into one string. But I don't see any code that is actually trying to so that.

Re: Capture error record in macro variable from a dataset

Kurt_Bremser — Thu, 03 Oct 2019 13:14:52 GMT

After fixing the missing comma and removing a wrong %end, it works without errors:

%macro IFR_WRITE_ERROR(
			error_code =
			, error_text =
			, flt_id =
			,flt_rec_txt =		
			);

	data protocol;
		ERROR_CODE = &error_code.;
		TEXT = "&error_text.";
		flt_id=&flt_id.;
		flt_rec_txt="&flt_rec_txt.";
	run;

%mend IFR_WRITE_ERROR;

%macro IFR_CHECK_KEYS;

	proc sql noprint;
		select count(*)
			into :dup_columns
				from dups
		;
	quit;

	%if &dup_columns. > 0 %then
		%do;
			%IFR_WRITE_ERROR(
				error_code = 100
				, error_text = Validation key - unique key violated
				,flt_id = 128
				,flt_rec_txt=dups
				);
			%let processing_code = 1;
		%end;

%mend IFR_CHECK_KEYS;

data dups;
x = 1;
output;
output;
run;

%IFR_CHECK_KEYS;

Re: Capture error record in macro variable from a dataset

David_Billa — Thu, 03 Oct 2019 13:47:32 GMT

What is the use of the macro variable processing_code?

Re: Capture error record in macro variable from a dataset

David_Billa — Thu, 03 Oct 2019 13:49:32 GMT

Because I'm not certain how to implement that logic in my program.

Re: Capture error record in macro variable from a dataset

David_Billa — Thu, 03 Oct 2019 14:34:53 GMT

What if the value of X has more than one values? Or

The dups dataset has multiple variables and values?

Re: Capture error record in macro variable from a dataset

Tom — Thu, 03 Oct 2019 14:48:45 GMT

Please post your data in a usable format.

data have;
  infile cards dsd dlm='|' truncover;
  input FLT_LAYR_NM :$10. FYP_NM :$30. _TEAM_NM :$30. ;
cards;
DIS|Consistencycheck|Business
DID|Rangecheck|Client
;
data expect;
  infile cards dsd dlm='|' truncover;
  input error_code error_text :$100. flt_id flt_rec_txt :$100. ;
cards;
100| |128|DIS Consistencycheck Business DID Rangecheck Client
;

Not clear that there is any need for macro logic at all for this problem.

data want;
  length error_code 8 error_text $100 flt_id 8 flt_rec_txt $100 ;
  keep error_code -- flt_rec_txt;
  error_code=100;
  error_text=' ';
  flt_id=128;
  do until(eof);
    set have end=eof;
    flt_rec_txt=catx(' ',flt_rec_txt,flt_layr_nm,fyp_nm,_team_nm);
  end;
run;

       error_    error_
Obs     code      text     flt_id                        flt_rec_txt

 1       100                 128     DIS Consistencycheck Business DID Rangecheck Client

Re: Capture error record in macro variable from a dataset

David_Billa — Thu, 03 Oct 2019 15:11:32 GMT

There are too many checks in the form of macro program as per my initial
post. I just placed only the summarzied code for one check.

Therefore I told that I want to tweak the code only in macro

Re: Capture error record in macro variable from a dataset

David_Billa — Fri, 04 Oct 2019 05:58:21 GMT

After executing your code, I see the Output as follows. Instead of populating the values of the variables from 'dups' dataset in the last variable 'flt_rec_txt' it has written the text 'dups'. Could you please help me resolve this issue here?

ERROR_CODE	TEXT	flt_id	flt_rec_txt
100	Validation key - unique key violated	128	dups

Whereas the desired Output is

ERROR_CODE	TEXT	flt_id	flt_rec_txt
100	Validation key - unique key violated	128	1

Re: Capture error record in macro variable from a dataset

David_Billa — Fri, 04 Oct 2019 10:29:52 GMT

If you want to handle the logic using macros, then how can you do that? Assume there are several HAVE datasets with multiple variables.

Re: Capture error record in macro variable from a dataset

Tom — Fri, 04 Oct 2019 13:35:48 GMT

@David_Billa wrote:

If you want to handle the logic using macros, then how can you do that? Assume there are several HAVE datasets with multiple variables.

It sounds like you want to process each one and APPEND the results to some aggregate dataset.

If the structure of the files is the same then the macro it probably very simple and just need to know the name of the current dataset that needs to be processed.

%macro do_one(dsname);

/* logic to process one file */
data this_one ;
  set &dsname ;
 ... 
run;

/* Aggregate to the master list */
proc append base=all_of_them data=this_one force;
run;

%mend do_one;

Re: Capture error record in macro variable from a dataset

David_Billa — Fri, 04 Oct 2019 13:38:32 GMT

Problem is structure of the files are different. Then how to tackle it?

Re: Capture error record in macro variable from a dataset

Tom — Fri, 04 Oct 2019 13:40:06 GMT

@David_Billa wrote:
Problem is structure of the files are different. Then how to tackle it?

No idea. How are they different? Why are they different? What is it want out of them?

Perhaps you don't actually want any of the data out of the files, just to create some information about the files. Like how many observations they contain.