topic Re: Merging 3 datasets in SAS Programming

Merging 3 datasets

geronimo44 — Mon, 21 Sep 2015 17:09:19 GMT

Hi,

I want to merge 3 datasets dataA, dataB and dataC, keeping records only found in dataA or dataB and exclude anything found in dataC. What approch would be more effective? SET or MERGE? I'm not able to run code at the moment so as to test the effeciency. Thus, I am trying to figure out through support and Help the merging funcions. I could do it with two datasets but still struggle with keeping only records found in A and B. I'd appreciate your ideas

Re: Merging 3 datasets

Reeza — Mon, 21 Sep 2015 17:12:20 GMT

Are you merging (side by side) or appending(adding more rows)?

Re: Merging 3 datasets

geronimo44 — Mon, 21 Sep 2015 17:24:43 GMT

It's an exercise and the description doesn't give more details on that. In my understanding it is asked to merge the datasets side by side

Re: Merging 3 datasets

Reeza — Mon, 21 Sep 2015 17:31:37 GMT

Well, set appends data sets so you're looking for a MERGE then. Look at the dataset IN option,example 4.
http://www.ats.ucla.edu/stat/sas/modules/merge.htm

Re: Merging 3 datasets

geronimo44 — Mon, 21 Sep 2015 17:52:42 GMT

So, something like the following would be correct? inputs and numbers are totally random

Data dataA

input x, y, z;

cards;

5, 10, 15

55, 24, 29

run;

proc sort data=dataA;

by x;

run;

Data dataB

input x, y, z;

cards;

105, 110, 150

run;

proc sort data=dataB;

by x;

run;

Data dataB

input x, y, z;

cards;

665, 1110, 1150

780, 455, 447

159, 668, 31

;

run;

proc sort data=dataC

by x;

run;

Data Merged_data

Merge dataA(IN fromdataA) dataB(IN fromdataB) dataC

BY x;

fromdataA=fromdataB;

run;

Re: Merging 3 datasets

Reeza — Mon, 21 Sep 2015 19:37:08 GMT

Close, but no cigar

Double check your output, especially if you have duplicate X's in any of your datasets.

Data Merged_data
Merge dataA(IN=fromdataA) 
           dataB(IN=fromdataB) 
           dataC(in=fromdataC)
BY x;

*Create indicator variables to show in which dataset the observation is present;
SourceA=fromdataA;
SourceB=fromdataB;
SourceC=fromdataC;

*keep only from A and B;
if fromdataA AND fromdataB;

run;