Solved
Contributor
Posts: 20

# Calculations with multiple observations

Hi All,

I have data like below

Date          B            C        D          E

20141231 193.83  93.04  129.98  96.15
20150102 194.41  93.02  129.95  95.03
20150105 188.34  90.56  129.05  93.5
20150106 184.53  88.63  127.53  92.95
20150107 187.28  90.3    129.51  94.87
20150108 190.27  91.58  131.8   97.06
20150109 187.35  90.42  131.54  95.99
20150112 185.07  89.5    30.87    95.86
20150113 184.93  89.23  131.17  95.01

I need to calcute Score = B1*B2+C1*C2+D1*D2+E1*E2

Accepted Solutions
Solution
‎12-23-2015 11:27 AM
Posts: 1,147

## Re: Calculations with multiple observations

[ Edited ]

Try to get the next record adjacent to the current record and then use the formula.

``````data have;
input Date          B1            C1        D1          E1;
cards;
20141231 193.83  93.04  129.98  96.15
20150102 194.41  93.02  129.95  95.03
20150105 188.34  90.56  129.05  93.5
20150106 184.53  88.63  127.53  92.95
20150107 187.28  90.3    129.51  94.87
20150108 190.27  91.58  131.8   97.06
20150109 187.35  90.42  131.54  95.99
20150112 185.07  89.5    30.87    95.86
20150113 184.93  89.23  131.17  95.01
;

data want;
obss=_n_+1;
set have end=last; if obss<=nobs then set have(keep=b1 c1 d1 e1 rename=(b1=b2 c1=c2 d1=d2 e1=e2)) end=last nobs=nobs point=obss;
if not last then score=sum(B1*B2,C1*C2,D1*D2,E1*E2);
else do;
score=.;
call missing(b2,c2,d2,e2) ;
end;
run;
``````

Thanks,
Jag

All Replies
Super User
Posts: 23,771

## Re: Calculations with multiple observations

``````Your formula is correct and will work directly in SAS, by adding a semi colon.

What issue are you having?

Score = B1*B2+C1*C2+D1*D2+E1*E2;``````
Contributor
Posts: 20

## Re: Calculations with multiple observations

Reeza,

I need to multiply first obervation with second and second with third and thrid with fourth.etc.and summing all the reult.

Super User
Posts: 23,771

## Re: Calculations with multiple observations

Post the expected output from the data above. Are you looking forward or looking back - does it matter? Looking back is easier, you can use the lag function.

score=B*lag1(b)+c*lag(c) + d1*lag(d) +e*lag1(e);
Solution
‎12-23-2015 11:27 AM
Posts: 1,147

## Re: Calculations with multiple observations

[ Edited ]

Try to get the next record adjacent to the current record and then use the formula.

``````data have;
input Date          B1            C1        D1          E1;
cards;
20141231 193.83  93.04  129.98  96.15
20150102 194.41  93.02  129.95  95.03
20150105 188.34  90.56  129.05  93.5
20150106 184.53  88.63  127.53  92.95
20150107 187.28  90.3    129.51  94.87
20150108 190.27  91.58  131.8   97.06
20150109 187.35  90.42  131.54  95.99
20150112 185.07  89.5    30.87    95.86
20150113 184.93  89.23  131.17  95.01
;

data want;
obss=_n_+1;
set have end=last; if obss<=nobs then set have(keep=b1 c1 d1 e1 rename=(b1=b2 c1=c2 d1=d2 e1=e2)) end=last nobs=nobs point=obss;
if not last then score=sum(B1*B2,C1*C2,D1*D2,E1*E2);
else do;
score=.;
call missing(b2,c2,d2,e2) ;
end;
run;
``````

Thanks,
Jag
Contributor
Posts: 20

## Re: Calculations with multiple observations

Thanks Reeza and Jagadishkatam, Its working now. Hooray!

🔒 This topic is solved and locked.

Discussion stats
• 5 replies
• 268 views
• 2 likes
• 3 in conversation