Extract part of an observertion

[desireatem]

Hello , I wish to create a variable "M" that extract the observations marked "bold" from days

Obs	days
1	20150320
2	20140609
3	20140807
4	20140828
5	20150418
6	20150602
7	20150625
8	20141029
9	20150220
10	20150407

Thank you!!!

[Jagadishkatam · Accepted solution]

Hoping that the days variable is a character variable

data want;
set have;
m=substr(strip(days),5,2);
run;

[desireatem]

Thank you!!!!

[RW9]

Whilst @Jagadishkatam is quite correct, that with the example you have given in a string variable, you can substr out various parts, I would also point out two things.  Firstly that is a date field, and you would be best off treating it as a date.  Second as that is a string field you could have anything in it, I have seen:

YY0401 or 120405

The substr would not work in these cases, unless of course you do various checks before hand.  So when working with dates I always have the two components, the string value for display purposes to show it as entered, and a numeric version for processing.  You could for instance do:

mnth=month(input(days,anydtdte.));

Though why you would call a date variable days is beyond me.

[desireatem]

who cares what name I give to my variables. I just wanted to extract some number from it and some showed me, what is your issue with it.

[desireatem]

This is it, working well, what is your issue with trying to be coolest:

data have;
input obs days;
cards;

1 20150320
2 20140609
3 20140807
4 20140828
5 20150418
;
run;
data want;
set have;
m=substr(strip(days),5,2);
run;


proc print data=want;
run;
Obs obs days m
1 1 20150320 03
2 2 20140609 06
3 3 20140807 08
4 4 20140828 08
5 5 20150418 04