Why people multiple by 100 to year(date) from CRSP?

[Phil_NZ]

Hi SAS Users,

This post is of my curiosity when reading a published code for a paper relating to finance (data from CRSP)

I do not have chances to access the dataset so what I guess just a guess.

The dataset has a variable name "date". I guess this variable have the informat date9.

I saw the author does one calculation surprising me, he calculates:

ym=year(Date)*100 + month(Date);

I am not sure if you deal with "date", whether you have ever seen any application of the above code?

I am sorry for the ambiguous information (because I just have the code, no more than that), but I hope that I can hear something from your experience.

Warmest regards.

Updated: I read closer to the later code, it seems that this code is to calculate the month.

[yabwon · Accepted solution]

Run this code:

data _null_;
  date = '08mar2021'd; /*today();*/
  format date date9.;

  y = year(Date);
  m = month(Date);

  ym=year(Date)*100 + month(Date); /* numerical value of "yeramonth", e.g. 202103 */

  put (_ALL_) (=/);
run;

Purpose of that line is to create numerical variable with "yearmonth" value, e.g for "08mar2021'd you will get 202103

Bart

[Kurt_Bremser]

Personally, I would create a character variable with

put(date,yymmn6.)

[yabwon]

Character is "1 function" and 6 bytes to store.

Numeric is "2 functions" but only 4 bytes to store,

data test;
  format date date9.;
  length ym 4;

  do date = '01jan1600'd, today(), '31dec9999'd;
    ym=year(Date)*100 + month(Date); 
    put (_ALL_) (=/);
    output;
  end;
run;
proc print;
run;

Possible question is: "Is comparison or sorting of numeric faster than character?"

Bart