topic Re: Hash object and find similiarty in SAS Programming

Hash object and find similiarty

erdem_ustun — Sat, 16 Apr 2016 14:14:40 GMT

Hi,

I have a data as it below and lots of records.

id,sid,ssn_1,ssn_2,ssn_3,fname,mname,is_same
1,1,101,245,655,aa,bb,base
2,1,101,202,655,AAA,bb,
3,1,103,200,657,CCC,bb,
4,1,103,245,655,aa,DDE,
5,1,104,245,699,BBB,MKE,
6,2,303,490,1310,Maa,bb,base
7,2,303,404,1310,FFF,bb,
8,2,309,400,1314,KKK,bb,
9,2,309,490,1310,Maa,DDE,
10,3,312,735,2097,BBB,TME,base
11,3,303,735,1965,aa,bb,
12,4,1212,808,2620,LARE,KEAN,base
13,4,1236,800,2628,CCC,bb,
14,4,1236,808,2620,LARE,KAANE,
15,4,1212,980,2796,BBB,KEAN,
..,..,..,..,..,..,..,..

I want to do is to compare with hash object the records and fields by same sid .İf it is same all fields .

All variables are the same ' put YYYYY' to is_same.if only first (ssn_1) and third(ssn_3) fields are the same put YNYNN' to is_same.

is it possible to use nested hash object?

The reference point for each sid where result='base'

for example first loop is

reference_1: id=1 and sid=1 and ssn_1=101 and ssn_2=245 and ssn_3=655 and fname=aa and sname=bb

compare_1: id=2 and sid=1 and ssn_1=101 and ssn_2=202 and ssn_3=655 and fname=AAA and sname=bb

result_1 :for ssn_1 is same and dor ssn_2 is not the same and for ssn_3 is same and for fname is not the same and for sname is same then result=YNYNY

....loop

if sid changed, the reference point will change.

How do I resolve this situation.

I want to result for sid=1

id,sid,ssn_1,ssn_2,ssn_3,fname,mname,is_same
1,1,101,245,655,aa,bb,base
2,1,101,202,655,AAA,bb,YNYNY
3,1,103,200,657,CCC,bb,NNNNN
4,1,103,245,655,aa,DDE,NYYYN
5,1,104,245,699,BBB,MKE,NYNNN
..,..,..,..,..,..,..,..

is it possible to solve to nested and conditional hash object for in order to be fast and effective.

best regards..

Re: Hash object and find similiarty

Ksharp — Sun, 17 Apr 2016 02:41:39 GMT

Hash Object has a method EQUAL() to judge whether two Hash Object is equal .But the way I looked, you don't need this method .


data have;
infile cards truncover dlm=',';
input id sid ssn_1 ssn_2 ssn_3 fname $ mname $ is_same $;
cards;
1,1,101,245,655,aa,bb,base
2,1,101,202,655,AAA,bb,
3,1,103,200,657,CCC,bb,
4,1,103,245,655,aa,DDE,
5,1,104,245,699,BBB,MKE,
6,2,303,490,1310,Maa,bb,base
7,2,303,404,1310,FFF,bb,
8,2,309,400,1314,KKK,bb,
9,2,309,490,1310,Maa,DDE,
10,3,312,735,2097,BBB,TME,base
11,3,303,735,1965,aa,bb,
12,4,1212,808,2620,LARE,KEAN,base
13,4,1236,800,2628,CCC,bb,
14,4,1236,808,2620,LARE,KAANE,
15,4,1212,980,2796,BBB,KEAN,
;
run;
data want;
 set have;
 by sid;
 if first.sid then do;
  _ssn_1=ssn_1; _ssn_2=ssn_2; _ssn_3=ssn_3; _fname=fname; _mname=mname;
 end;
 else do;
  is_same=cats(ifc(ssn_1=_ssn_1,'Y','N'),
               ifc(ssn_2=_ssn_2 ,'Y','N'),
               ifc(ssn_3=_ssn_3 ,'Y','N'),
               ifc(fname=_fname ,'Y','N'),
               ifc(mname=_mname,'Y','N'));
 end;
 retain _ssn_1 _ssn_2 _ssn_3 _fname _mname;
 drop _:;
run;

Re: Hash object and find similiarty

erdem_ustun — Mon, 18 Apr 2016 07:12:02 GMT

Thank you very much for your answer.

I wanted to do with has object for the speed and efficiency.

I understand that you said there hash object Where equality occurs of this method.

Thank you for your explanation.

The code you have written comparing the next record.I did an insert for code and results occurred exactly what I expected.

if first.sid and is_same='base' then do;

In the final code below ;

data want;
 set have;
 by sid;
 if first.sid and is_same='base' then do;
  _ssn_1=ssn_1; _ssn_2=ssn_2; _ssn_3=ssn_3; _fname=fname; _mname=mname;
 end;
 else do;
  is_same=cats(ifc(ssn_1=_ssn_1,'Y','N'),
               ifc(ssn_2=_ssn_2 ,'Y','N'),
               ifc(ssn_3=_ssn_3 ,'Y','N'),
               ifc(fname=_fname ,'Y','N'),
               ifc(mname=_mname,'Y','N'));
 end;
 retain _ssn_1 _ssn_2 _ssn_3 _fname _mname;
 drop _:;
run;

best regards..