topic Re: Removing select cases from a dataset in SAS Procedures

Removing select cases from a dataset

Angi — Thu, 05 Feb 2015 01:00:55 GMT

Hi,

I have two datasets in long format that have variables “ID” and “incident_number” in common. Each person has a unique ID number but two or more people can have the same incident_number if they violate school policy together (for example, both of them engaged in underage drinking at the same party; so the underage drinking at the party would count as one incident).

If DatasetB doesn’t have a matching ID and incident number with DatasetA, I want to delete unmatched cases in DatasetB. So, I want to select cases based on a combination of ID and incident number.

Is there a way to modify DatasetB in this way?

Have DatasetA:

ID	Incident_number	VarA	VarB
1	56789	Abcd	1232
1	56789	Erfsdjdgr	94847
2	64908	Fakjsdf;asie	5342
2	56789	Sajfldkjsa;ljf	54698734
3	78621	Kjdafjdh	0589576
4	78621	afsdfasdkljfa	1154
4	78721	Fsafd	987888

Have DatasetB:

ID	Incident_number	VarX	VarY
1	56789	56789	galdjfoia
1	56789	56789	sadjfakhd
1	56818	56818	weoiatlk
2	56789	124363	Abjdkafh
2	64826	5463	Giagaojdfl
3	78629	1545744654	Ajfu
3	78799	1265841	Fkajhdfajh
4	78621	9378464	Nadfaei
4	79725	5486513541	Lirueoi
9	79102	23756	Nann
12	80215	987543	Ueyrf

Want new_DatasetB

ID	Incident	VarX	VarY
1	56789	56789	galdjfoia
1	56789	56789	sadjfakhd
2	56789	124363	Abjdkafh
4	78621	9378464	Nadfaei

Thank you for your help.

Re: Removing select cases from a dataset

Miracle — Thu, 05 Feb 2015 03:51:44 GMT

Try this,

proc sort data=a; by ID Incident_number; run;

proc sort data=b; by ID Incident_number; run;

data want; merge a(in=a) b(in=b); by ID Incident_number; if a=b; drop VarA VarB; run;

proc print noobs; run;

Re: Removing select cases from a dataset

Angi — Thu, 05 Feb 2015 05:37:10 GMT

Thank you K C Wong. Is there any way of achieving the same result without merging the datasets?

Re: Removing select cases from a dataset

Ksharp — Thu, 05 Feb 2015 07:47:18 GMT

Hash Table ?

data DatasetA;
input 
ID
Incident_number
VarA $
VarB ;
cards4;
1
56789
Abcd
1232
1
56789
Erfsdjdgr
94847
2
64908
Fakjsdf;asie
5342
2
56789
Sajfldkjsa;ljf
54698734
3
78621
Kjdafjdh
0589576
4
78621
afsdfasdkljfa
1154
4
78721
Fsafd
987888
;;;;
run;
data DatasetB;
input 
ID
Incident_number
VarX 
VarY $;
cards4;
1
56789
56789
galdjfoia
1
56789
56789
sadjfakhd
1
56818
56818
weoiatlk
2
56789
124363
Abjdkafh
2
64826
5463
Giagaojdfl
3
78629
1545744654
Ajfu
3
78799
1265841
Fkajhdfajh
4
78621
9378464
Nadfaei
4
79725
5486513541
Lirueoi
9
79102
23756
Nann
12
80215
987543
Ueyrf
;;;;
run;

data want;
 if _n_ eq 1 then do;
  declare hash h(dataset:'DatasetA');
  h.definekey('ID','Incident_number');
  h.definedone();
 end;
set DatasetB;
if h.check()=0;
run;

Xia Keshan

Re: Removing select cases from a dataset

Angi — Sat, 07 Feb 2015 23:44:55 GMT

Thank you very much xia keshan!