The answer is actually more complex than you would think.  It depends in part on whether you have a one-to-one match or a many-to-one match.  If you are interested in reading and learning, here is a link to an old paper than explains the process:   https://www.lexjansen.com/nesug/nesug99/ad/ad155.pdf   ... View more

The easiest way: include the semicolon when replacing RESULT1. The 3rd parameter in CATX would become: 'Hello world:' Do you have some conditions that would make this a bad idea? ... View more

We can't see your data.  We can't see what is in your macro variables.  So we are guessing.  But here is my guess.  SImplify this part of your query: and (c.srcstcd not in ('5','6') or c.srcstcd is null) and (d.pur_id not in ('0001234','0005678') or d.pur_id is null) Even if the query were working successfully, you would get the same result by using: and (c.srcstcd not in ('5','6') ) and (d.pur_id not in ('0001234','0005678')) After all, if c.srcstcd is null, the first condition about it not being '5' or '6' would capture the observation.  So even if this doesn't turn out to be the source of your error, at least you can simplify the program without changing the meaning of the query. ... View more

As others have indicated, the statements will run without error.  However, there is a flaw.  If those statements are creating a new variable named PRODUCT, it will be defined as two characters long.  The first mention of PRODUCT (= 'EC') defines its length.  So the values you will get are "EC", "bc", and "ot".  The fix is simple:  add your own LENGTH statement before SAS sets the length for you: data test; set test2; length product $ 7; if card_type = 'F' then product = 'EC'; else if customer = 'C' then product = 'bc_card'; else if card_type in ('dg', 'dp', 'bg') then product = 'others'; run; The length of 7 allows room for the longest value being assigned ("bc_card"). ... View more

Good job running the program I requested.  The results are a bit difficult to read.  Do they appear the same before you posted them here?     You have two tables, with the same numbers in them.  That is good.  The idea now is to take find the match between the displayed values, such as "Other tranqualizers" and the codes that the data uses to represent those values.  Compare the two tables, and see if it's obvious which codes match which categories.    If the two tables can't be compared in that way, we'll devise another way.  But for now see if it is obvious to you how to match up the codes and the categories. ... View more

This type of error can easily occur when you import data. You don't really know what is in the data, and so you need to explore to find out. Run this test as the initial exploratory step: proc freq data=have; tables sub1; run; proc freq data=have; tables sub1; format sub1; run; Of course you would not use the word "have" as your data set name. Use the actual name of your imported data set. ... View more

The log here shows that you ran PROC FORMAT successfully.  Having done that, the next step would be to actually apply the format you have created.  For example: proc freq data=work.tedsa_puf_2021; tables sub1 * gender; format sub1 $mycat.; run; Try that, and see if it starts to make sense. ... View more

Here's an approach that assumes SUB1 is a character variable. If that's not the case, we'll need more details about what is actually in your data. Don't change the data. Instead, create a format that tells PROC FREQ how to combine categories. For example you might code: proc format; value $mycat "Alcohol", "Barbiturates", "Other Tranquilizers" = "Combined ABT" : run; Type in the full list of any values that should be combined. Call the combined category anything you would like. "Combined ABT" is just an arbitrary name that i invented. Spelling is crucial. Capitalization is crucial. Once this is ready, group the data by adding the following statement after each TABLE statement: format sub1 $mycat ; ... View more

What does PROC CONTENTS tell you about QN1 and QN2? Are they character or are they numeric? ... View more

A few points to begin.   PROC FREQ will probably not be part of the solution.  It doesn't compute the statistics you are hoping to display.   PROC TABULATE will do the job (and likely PROC REPORT as well, but I'm not as familiar with it).  And you won't need to manipulate the data ahead of time.   What procedure are you hoping to use to get p-values?  What do the p-values represent?  There's no way to include p-values in a program until you can address those questions.   Here's a shell of a program to get you going:   proc tabulate data=want; class country; var age height; tables (age height) * (n mean median std q1 q3 min max), country; run; ... View more

The values like < .0001 are already numeric. They are printing using a format that includes a less than sign for small values. One choice you have is to remove the format before printing: format var; Another choice you have is to change the data: if var < .0001 then var = .0001; Does that make sense? ... View more

The coalesce function would work, as @Kurt_Bremser suggested.  You could also write an IF/THEN statement as you suggested, along these lines: IF ULINKENDDT <= .Z THEN ULINKENDDT = '31dec2023'd; ... View more

Loading an observation into an array is trivial.  So just to make sure we are talking apples to apples, here are some points to consider.   Do you expect the array to be saved somehow?  In almost all cases, array definitions disappear as soon as the DATA step ends.   Do you really have just 3 variables, or could there be hundreds (or thousands)?  How will you determine the number of variables?   Do you need the array values to be available when you process the first observation in the data set?   When the program reads the 3rd observation, you may be replacing the contents of the array.  Does that sound OK?   The already-asked question of what do you plan to do from here likely becomes relevant. ... View more

Does this string represent October 2 or February 10, or do you want SAS to make that choice for you? ... View more