topic Re: How to match tables based on different criteria in SAS Programming

How to match tables based on different criteria

cindyforest7 — Mon, 24 Mar 2014 20:04:12 GMT

Hi,

I have two identical tables, and I want to match them together based on Age_grp, sex, and any_comorbidity.

Members are connected if they have same age_grp, or same sex, or same any_comorbidity. If two memebrs have three things in common, then the weight of the connection is 3; if two members have two things in common, then the weight of the connection is 2, and if two members only have one thing in common, then the weight should be 1.

A member cannot be connected to himself, and only connection with weight of at least of 1 should be shown.

Table 1

Member_id Age_grp sex any_comorbidity

a 1 F Y

b 1 M Y

c 2 M N

d 1 M N

e 3 F N

f 5 M Y

g 2 M Y

h 3 F Y

Table 2

Member_id_match Age_grp sex any_comorbidity

a 1 F Y

b 1 M Y

c 2 M N

d 1 M N

e 3 F N

f 5 M Y

g 2 M Y

h 3 F Y

I want a final table looks like this

Member_id Member_id_match weight

a b 2

a d 1

a e 1

a f 1

a g 1

a h 2

....

Does any one have a solution here? Much appreciated!!!

Re: How to match tables based on different criteria

PGStats — Mon, 24 Mar 2014 20:22:03 GMT

Relatively simple with SQL :

data have;

input id $ Age_grp sex $ any_c $;

datalines;

a 1 F Y

b 1 M Y

c 2 M N

d 1 M N

e 3 F N

f 5 M Y

g 2 M Y

h 3 F Y

;

proc sql;

create table match as

select

a.id,

b.id as match_id,

(a.age_grp=b.age_grp) + (a.sex=b.sex) + (a.any_c=b.any_c) as weight

from

have as a cross join

have as b

where a.id < b.id and calculated weight > 0;

quit;

Re: How to match tables based on different criteria

slchen — Mon, 24 Mar 2014 22:24:32 GMT

data want;

set have;

do point=1 to nobs;

set have (rename=(id=_id age_grp=_age_grp sex=_sex any_c=_any_c)) point=point nobs=nobs;

if point>_n_ and id^=_id then do;

weigth=sum(age_grp=_age_grp,sex=_sex,any_c=_any_c);

if weigth>0 then output;

end;

keep id _id weigth;

run;

Shenglin