topic Want a character operand, not a numeric, but is not evaluating in SAS Data Science

Want a character operand, not a numeric, but is not evaluating

Re: Want a character operand, not a numeric, but is not evaluating

Re: Want a character operand, not a numeric, but is not evaluating

Re: Want a character operand, not a numeric, but is not evaluating

Re: Want a character operand, not a numeric, but is not evaluating

Re: Want a character operand, not a numeric, but is not evaluating

zhou_larry — Thu, 16 Apr 2015 19:31:26 GMT

/*format begin_date end_date date9.;*/

%LET start = %sysevalf(begin_date);

%LET end = %sysevalf(end_date);

I actually want to pass the date from above, 26JUN2016 to the %start variable and 03JUL2016 to the %end variable and have been trying different %EVALs to get it to evaluate what is begin_date and what is end_date

The issue is that if I just plainly put in

%LET start = begin_date;

it only passes along the text "begin_date" and does not solve for the what begin_date is assigned to.

and if I use %EVAL, then the macro looks for a numeric whereas I actually want the date (or text) of "26JUN2016" is equal to &start

Reeza — Thu, 16 Apr 2015 19:45:13 GMT

Is that in a data step?

If so to create a macro variable in a data step you need to use call symputx to create the macro variable not a %let statement.

WendyCzika — Thu, 16 Apr 2015 20:09:35 GMT

Yes as Reeza says, you would need to do something like the following inside the DATA step:

Then after the DATA step, the macro variables start and end are available to use with the actual dates in them.

zhou_larry — Thu, 16 Apr 2015 20:47:06 GMT

Indeed it is. So I changed it to:

call symputx("start", begin_date);

call symputx("end", end_date);

format begin_date end_date start end date9.;

%put &start;

%put &end;

and now it's passing through!!

but now the date is a SAS date, so '12APR2015' = 20190 which isn't ideal. and formatting it isn't exactly working

So the next data step is something like:

data work.xx;

keep x, y, z;

WHERE (&start) < datepart(order_date) <= (&end);

but now because it's a SAS date, it looks at 20190 < datepart(order_date) which comes up blank when it runs. How do I made the order_date into a SAS format or how do I make the 20190 pass as a date9. instead of a SAS date?

Reeza — Thu, 16 Apr 2015 23:40:40 GMT

Actually your WHERE condition should still resolve properly.

Your data step is incorrect the keep statement has comma's and there is no SET dataset.

...
...
ELSE IF '26JUN2016'd < todaydate <= '03JUL2016'd THEN DO;

begin_date = '26JUN2016'd;

end_date='03JUL2016'd;

Tom — Fri, 17 Apr 2015 12:17:15 GMT

If you want the macro variables to contain values that would be valid to use in generating the logical expression

(&start < XX < &end)

where XX is a date expression then you have two choices.

1) Store the actual integer values that represent the start and end dates.

CALL SYMPUTX('START',put(begin_date,10.))

2) Store the value as a date literal.

CALL SYMPUTX('START',cats("'",put(begin_date,date9.)."'d"))

You could also create the macro variable with DATE9 formatted date strings and then add the ""d in the where clause.

CALL SYMPUTX('START',put(begin_date,date9.));

...

("&start"d < XX < "&end"d)