Proc Tabulate with Out=

[buechler66]

Hi. I'm using Proc Tabulate with the Out= option. In my data I keep 3 variables:

1. rule_order

2. ad_dt

3. n

Proc tabulate data=FinalData missing out=testdata (keep=rule_order ad_dt n) ;
class Rule_Order ad_dt;
table rule_order*(ad_dt all)*n;
run;

My problem is that the AD_DT's that are missing have a value of . (missing) but my totals by Rule_Order also have a . (missing). In the data I would like to be able to recode the totals that have a . to be 'Total', but I don't see a way to do this.

I've attached a sample of what my output dataset Testdata looks like.

Any help would be much appreciated.

[Astounding]

By default, PROC TABULATE omits observations where any CLASS variable has a missing value.  You can have them included by adding the MISSING option.

[Reeza]

class Rule_Order ad_dt /Missing ;

[ballardw · Accepted solution]

Look closely at the output data. You will find a variable _type_. This, similar to proc Means and Summary, indicates the combinations of variables used to create the values for a row of the data set. Determine which level(s) of _type_ represent your summary, usually not too difficult unless you have many class variables and "All" clauses.

However you are likely to have serious issues about attempting to assign a value of "Total" to a NUMERIC variable (ie you can't).

Depending on complexity you might get away with a custom format to assign a value of 'Total' to missing.