BookmarkSubscribeRSS Feed
🔒 This topic is solved and locked. Need further help from the community? Please sign in and ask a new question.
devsas
Pyrite | Level 9

Hello everyone,

So, I have these two datasets as below-

 

data test1;
input ssn loanNO effDate ddmmyy10. program $ ;
cards;
1234 1 01/01/2013  no
1234 2 01/05/2013  no
1234 3 01/01/2014  yes
5678 1 01/01/2012  no
9101 1 01/05/2014  yes
9101 2 01/09/2013  no
;
run;

data test2;
input ssn fico date ddmmyy10. ;
cards;
1234 600 01/02/2012
1234 580 01/05/2012
1234 700 01/08/2012
1234 680 01/11/2012
1234 500 01/02/2013
1234 580 01/05/2013
1234 700 01/08/2013
1234 680 01/11/2013
1234 600 01/02/2014
1234 580 01/05/2014
1234 600 01/08/2014
1234 690 01/11/2014

;
run;

 

I need to join only those records from first datset test1 where program is yes and then get corresponding fico scores for that person for that year, previous year and following year. So, for example, for ssn 1234, yes was for 1 jan 2014 and the fico score during that period was 680 (fico scores are for 4 periods for each year, feb, may, aug and nov and in case we are looking for month other than these 4 months, we use the previous one-for ex, in this case nov 2013 was the last one, so we use that). Similarly fico score for previous year during that period was also 680 for nov 2012 and for following year we use nov 2014 which was 690.

 

Thanks a lot in advance.

1 ACCEPTED SOLUTION

Accepted Solutions
art297
Opal | Level 21

Hopefully, one of us will have correctly guessed the full requirements. I think the following is what you want. Note: I changed the fico scores so that we could identify where they came from. The 3 digit numbers I used represent the year for the first digit, and the next two represent the month. Those values aren't necessary for the code to run, but make it a lot easier to see which fico scores were used:

data test1;
  input ssn loanNO effDate ddmmyy10. program $ ;
  format effDate date9.;
  cards;
1234 1 01/01/2013  no
1234 2 01/05/2013  no
1234 3 01/01/2014  yes
5678 1 01/01/2012  no
9101 1 01/05/2014  yes
9101 2 01/09/2013  no
;
run;

data test2;
  input ssn fico date ddmmyy10. ;
  format date date9.;
  cards;
1234 202 01/02/2012
1234 205 01/05/2012
1234 208 01/08/2012
1234 211 01/11/2012
1234 302 01/02/2013
1234 305 01/05/2013
1234 308 01/08/2013
1234 311 01/11/2013
1234 402 01/02/2014
1234 405 01/05/2014
1234 408 01/08/2014
1234 411 01/11/2014
9101 202 01/02/2012
9101 205 01/05/2012
9101 208 01/08/2012
9101 211 01/11/2012
9101 302 01/02/2013
9101 305 01/05/2013
9101 308 01/08/2013
9101 311 01/11/2013
9101 402 01/02/2014
9101 405 01/05/2014
9101 408 01/08/2014
9101 411 01/11/2014
9101 502 01/02/2015
9101 505 01/05/2015
9101 508 01/08/2015
9101 511 01/11/2015
;

proc sql noprint;
  create table want as
    select a.*, b.fico,b.date,
           c.fico as fico2, d.fico as fico3
      from test1 (where=(program eq 'yes')) a
        left join
          test2 b
            on a.ssn eq b.ssn and
               (a.effdate eq b.date or
               intnx('month',a.effdate,-1,'b') eq b.date or
               intnx('month',a.effdate,-2,'b') eq b.date)
              left join
                test2 c
                  on a.ssn eq c.ssn and
                     intnx('year',b.date,-1,'s') eq c.date
                    left join
                      test2 d
                        on a.ssn eq d.ssn and
                           intnx('year',b.date,1,'s') eq d.date
     ;
quit;

Art, CEO, AnalystFinder.com

View solution in original post

6 REPLIES 6
Patrick
Opal | Level 21

@devsas

In case I understood your business logic correctly below should do:

 

proc sql feedback;
  create table want as
    select t1.*, t2.date, t2.fico
    from 
      test1 as t1
      left join
      test2 as t2
        on 
          t1.ssn=t2.ssn and
          t2.date between intnx('year',t1.effDate,-1,'b') and intnx('year',t1.effDate,+1,'e')
      where t1.program='yes'
      order by t1.ssn, t2.date
    ;
quit;

 

Result:

Capture.PNG

 

If the code doesn't return what you want then please post a "desired result" table showing us the output you want based on your sample data.

 

devsas
Pyrite | Level 9

sorry guys, should have been more clearer. Thanks for your codes, i guess my below updated data along with want dataset should make it more clear. Please use these numbers now, the data is not necessarily sorted with dates etc.

data test1;
input ssn loanNO effDate ddmmyy10. program $ ;
cards;
1234 1 01/01/2013  no
1234 2 01/05/2013  no
1234 3 01/01/2014  yes
5678 1 01/01/2012  no
9101 1 01/05/2014  yes
9101 2 01/09/2013  no
;
run;

data test2;
input ssn fico date ddmmyy10. ;
cards;
1234 600 01/02/2012
1234 580 01/05/2012
1234 700 01/08/2012
1234 680 01/11/2012
1234 500 01/02/2013
1234 580 01/05/2013
1234 700 01/08/2013
1234 680 01/11/2013
1234 600 01/02/2014
1234 580 01/05/2014
1234 600 01/08/2014
1234 680 01/11/2014
5678 600 01/02/2012
5678 580 01/05/2012
5678 700 01/08/2012
5678 680 01/11/2012
5678 500 01/02/2013
5678 580 01/05/2013
5678 700 01/08/2013
5678 680 01/11/2013
5678 600 01/02/2014
5678 580 01/05/2014
5678 600 01/08/2014
5678 680 01/11/2014
9101 600 01/02/2015
9101 690 01/05/2015
9101 670 01/08/2015
9101 680 01/11/2015
9101 550 01/02/2013
9101 580 01/05/2013
9101 700 01/08/2013
9101 680 01/11/2013
9101 600 01/02/2014
9101 680 01/05/2014
9101 700 01/08/2014
9101 680 01/11/2014


;
run;

data want;
input ssn loanno ficopresent ficoprevYr ficoNextYr;
cards;
1234 3 680 680 680
9101 1 680 580 690
;
run;
Shmuel
Garnet | Level 18

Try next, not tested code:

data test2a;
 set test2;
  by ssn;
     date_from = lag(date);
     if first.ssn then call missing(date_upto );
     date_upto = date - 1;
     if not missing(date_from) then output;
     drop date;
run;

proc sql;
  create table want as select t1.*, t2.fico
  from test1 as t1
  left join test2a as t2
  on t1.ssn = t2.ssn and
     t1.effdate between t2.date_from and t2.date_upto;
quit
art297
Opal | Level 21

Hopefully, one of us will have correctly guessed the full requirements. I think the following is what you want. Note: I changed the fico scores so that we could identify where they came from. The 3 digit numbers I used represent the year for the first digit, and the next two represent the month. Those values aren't necessary for the code to run, but make it a lot easier to see which fico scores were used:

data test1;
  input ssn loanNO effDate ddmmyy10. program $ ;
  format effDate date9.;
  cards;
1234 1 01/01/2013  no
1234 2 01/05/2013  no
1234 3 01/01/2014  yes
5678 1 01/01/2012  no
9101 1 01/05/2014  yes
9101 2 01/09/2013  no
;
run;

data test2;
  input ssn fico date ddmmyy10. ;
  format date date9.;
  cards;
1234 202 01/02/2012
1234 205 01/05/2012
1234 208 01/08/2012
1234 211 01/11/2012
1234 302 01/02/2013
1234 305 01/05/2013
1234 308 01/08/2013
1234 311 01/11/2013
1234 402 01/02/2014
1234 405 01/05/2014
1234 408 01/08/2014
1234 411 01/11/2014
9101 202 01/02/2012
9101 205 01/05/2012
9101 208 01/08/2012
9101 211 01/11/2012
9101 302 01/02/2013
9101 305 01/05/2013
9101 308 01/08/2013
9101 311 01/11/2013
9101 402 01/02/2014
9101 405 01/05/2014
9101 408 01/08/2014
9101 411 01/11/2014
9101 502 01/02/2015
9101 505 01/05/2015
9101 508 01/08/2015
9101 511 01/11/2015
;

proc sql noprint;
  create table want as
    select a.*, b.fico,b.date,
           c.fico as fico2, d.fico as fico3
      from test1 (where=(program eq 'yes')) a
        left join
          test2 b
            on a.ssn eq b.ssn and
               (a.effdate eq b.date or
               intnx('month',a.effdate,-1,'b') eq b.date or
               intnx('month',a.effdate,-2,'b') eq b.date)
              left join
                test2 c
                  on a.ssn eq c.ssn and
                     intnx('year',b.date,-1,'s') eq c.date
                    left join
                      test2 d
                        on a.ssn eq d.ssn and
                           intnx('year',b.date,1,'s') eq d.date
     ;
quit;

Art, CEO, AnalystFinder.com

devsas
Pyrite | Level 9

Yes, sir, it seems to work. I actually posted my updated data again. Thanks so much1

Ksharp
Super User

Assuming I understand what you mean.

 

data test1;
input ssn loanNO effDate ddmmyy10. program $ ;
if program='yes';
cards;
1234 1 01/01/2013  no
1234 2 01/05/2013  no
1234 3 01/01/2014  yes
5678 1 01/01/2012  no
9101 1 01/05/2014  yes
9101 2 01/09/2013  no
;
run;

data test2;
input ssn fico date ddmmyy10. ;
if month(date) in (2 5 8 11);
cards;
1234 600 01/02/2012
1234 580 01/05/2012
1234 700 01/08/2012
1234 680 01/11/2012
1234 500 01/02/2013
1234 580 01/05/2013
1234 700 01/08/2013
1234 680 01/11/2013
1234 600 01/02/2014
1234 580 01/05/2014
1234 600 01/08/2014
1234 690 01/11/2014

;
run;
data temp;
 set test1(in=ina rename=(effDate=date)) test2 ;
 by ssn date;
 retain prev prev_date;
 if first.ssn then call missing(prev,prev_date);
 if not missing(fico) then do;prev=fico;prev_date=date;end;
 next_date=intnx('year',prev_date,1,'s');
 if ina;
 drop fico;
 format date prev_date next_date ddmmyy10.;
run;
data want;
 merge temp(in=ina) test2(rename=(date=next_date fico=next));
 by ssn next_date;
 if ina;
run;

sas-innovate-2024.png

Join us for SAS Innovate April 16-19 at the Aria in Las Vegas. Bring the team and save big with our group pricing for a limited time only.

Pre-conference courses and tutorials are filling up fast and are always a sellout. Register today to reserve your seat.

 

Register now!

How to Concatenate Values

Learn how use the CAT functions in SAS to join values from multiple variables into a single value.

Find more tutorials on the SAS Users YouTube channel.

Click image to register for webinarClick image to register for webinar

Classroom Training Available!

Select SAS Training centers are offering in-person courses. View upcoming courses for:

View all other training opportunities.

Discussion stats
  • 6 replies
  • 2448 views
  • 1 like
  • 5 in conversation