Substr YYYY and MM from Sas date YYYYMM (edited)

[Cruise]

I have a SAS date in format of YYYYMM as shown in my viewtable below. This YYYYMM "dob" variable is the truncation of original 8 digit SAS date for confidentiality purpose. So, I don't know how it originally looked like.

The final result that I would like to have is:

I tried several different ways and some of my unsuccessful attempts are here: 

dob1 = put(dob, YYMMN6.); 
dob1 = input(dob, yymmn6.);
dy = substr(dob,5,4);
dm = substr(dob,7,2);

Jagadishkatam's suggested code below was resulted as shown:

dob1=put(dob,yymmn6.);
dy = year(dob1);
dm = month(dob1);

Result of above code:

I haven't tried substrn yet. I will soon.

This problem was previously discussed here. But it didn't work with my SAS9.4.

Any hints are appreciated.

[Jagadishkatam]

We could use the year() and month() functions as below

dy=year(dob);
dm=month(dob);

Note: At the moment the date seem not right like the DOB of 189005 when converted to date9 format it is 23Jun2477. so it is a strange date. but if it is only for example then it is fine.

[Cruise]

Hi Jagadishkatam,

The solution you suggested was resulted as shown below. Your observation using data9. is very interesting. it's not just for display. I really have to figure out this yyyymm date variable. Current yyyymm date I have is the result of a truncation of 8 digit SAS date, I think.

[Jagadishkatam]

you could try the

using the put(dob,yymmn6.) it does retrieve the year and month. I am not sure what is the exact issue you have.

[Reeza · Accepted solution]

What does 189005 mean in terms of a date? 

What part is the year, day and month? 

Try SUBSTRN

year = substrn(old_var, 1, 4);
month = substrn(old_var, 5, 2);

Or

Date_type =input( put(old_var, 6.), yymmn6.);
Year=year(date_type);
Month = month(date_type);

If this doesn't work post the exact code and log. 

[Cruise]

Your code did the trick to a month. But years are still in 1809 something range. I will post exact details when i get back to office. I have a feeling that yours is getting close.

[Cruise]

Hi Reeza,

I tried your first code with and without left.

/*Reeza1*/
data x1; set x;
year = substrn(left(dob), 1, 4);
month = substrn(left(dob), 5, 2);
run; 
/*Reeza2*/
data feb.gro_id1; set feb.gro_id;
Date_type =input( put(dob, 6.), yymmn6.);
Year=year(date_type);
Month = month(date_type);
run; 

The corresponding results in viewtable are:

Reeza1:

Reeza2:

[Astounding]

It would be helpful to have an example of the math.  For example, show what YYYY and MM look like, an original DOB, and what the final result would look like after subtracting them.

[Cruise]

just edited my original post as per your suggestion

[Astounding]

If I have the question right (and I'm not really sure about that), you are starting with A and looking to calculate B and C.

If that's the case, the formulas are straightforward (and now corrected):

B = int(A/100);

C = mod(A, 100);

Somehow I get the feeling that I'm leaving out a key piece of the question though.  I can't draw the link between this table and your original table.

[Cruise]

Hi Respected,

I am starting with below. Date is truncated for confidentiality concern. So I have this 6 digit SAS date.

I tried your code as:

year=substr(left(dob),1,4);
month=substr(left(dob),5,2);
B = int(year/100);
C = mod(year, 100);

That is resulted in:

Any alternatives?

Thanks

[Astounding]

Just mixing and matching names ... should be:

year = int(dob/100);

month = mod(dob, 100);

[Cruise]

I still have 1890, 1898 sort of years after this.

[Tom]

If you are getting DOB in the 1800's then either you have some really old people or the process that created that field did something wrong.  

Most likely it assumed the wrong century for a date string with a two digit year.  If you cannot correct that then you will need to do the best you can to fix the DOB after the fact.  If you have other information that could lead you believe that the 1890 is unreasonable but that 1990 is reasonable for that person then just add 100 years to the DOB.

[Reeza]

@Cruise Look at the first picture you posted. It looks like 1890. If it's not supposed to be 1890 what is it supposed to be. According to what I see, the code looks like it ran correctly. Either

1. Your dates are in some other format than you've specified. Did you import this data from Excel, from a date field

2. Your dates are wrong

3. Dates are correct and you have some really old people.