topic Re: SAS append/merge datasets in SAS Programming

SAS append/merge datasets

jeremy4 — Thu, 01 Aug 2019 09:40:08 GMT

Hi,

I have two datasets - ar.dataset_2018 and ar.dataset_2019. Please can someone finish off the code below (highlighted in red/whether set is the correct method) that would combine the two datasets (screenshot below shows an extract of the individual 2018 and 2019 datasets and the third table is an example of what I would like the merged dataset to look like), with the new dataset only containing records between 1st January 2018 and 31st May 2019 (which I have included in the code I have written so far - where '01Jan2018'd <= month <= '31May2019'd; )? Thanks!

Code

data datasets_2018_2019;

set ar.dataset_2018 ar.dataset_2019; /*What code to use to merge the two datasets, as shown by the screenshot?*/

where '01Jan2018'd <= month <= '31May2019'd;

run;

Re: SAS append/merge datasets

KachiM — Thu, 01 Aug 2019 10:21:07 GMT

Not Tested.

Try

set ar.dataset_2018 ar.dataset_2019;

BY MONTH ID;

Re: SAS append/merge datasets

jeremy4 — Thu, 01 Aug 2019 10:24:15 GMT

Thanks for your reply. Would the code be:

data datasets_2018_2019;
set ar.dataset_2018 ar.dataset_2019;
by month ID;
where '01Jan2018'd <= month <= '31May2019'd;
run;

Re: SAS append/merge datasets

KachiM — Thu, 01 Aug 2019 10:25:16 GMT

YES

Re: SAS append/merge datasets

Astounding — Thu, 01 Aug 2019 12:25:58 GMT

It appears that you are taking ALL the observations from your first data set, followed by ALL the observations from the second data set. If that's the case, you don't need a WHERE statement A simple DATA step would be sufficient:

data datasets_2018_2019;

set ar_dataset_2018 ar_dataset_2019;

run;

Note that using a WHERE statement assumes that the MONTH variable is actually a SAS date representing a particular day, and not just a month and year without a day.

Another possibility ... this would run faster but would use the name AR_DATASET_2018 to hold the combined version of the data:

proc append data=ar_dataset_2019 base=ar_dataset_2018;

run;

In the case of such small data sets, it seems like speeding up the program is not valuable, however.