topic Re: Proc compare to compare the data sets in New SAS User

Proc compare to compare the data sets

Sandeep77 — Thu, 21 Jul 2022 12:32:00 GMT

Hi Experts,

I want to compare two huge data set with lot of rows and columns. With proc compare can I compare the entire data set or I have to select few columns and then compare? Also, is there a link where I can check how to write the proc compare statement?

Re: Proc compare to compare the data sets

Quentin — Thu, 21 Jul 2022 12:58:40 GMT

Yes, PROC COMPARE can compare entire datasets.

See the documentation:

https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/proc/n1nwxbchh5hpu1n1h28kmici2awd.htm

proc compare base=dataset1 compare=dataset2;
  id PatientID;  *Almost always you want an ID statement, otherwise it will compare by row;
run;

Re: Proc compare to compare the data sets

kleelasiva9 — Thu, 21 Jul 2022 13:00:43 GMT

Q: With proc compare can I compare the entire data set or I have to select few columns and then compare?

Ans: Both, but by default proc compare compares entire dataset. Using Var statement restricts the comparison of the values of variables to the ones named in the VAR statement.

Q: Also, is there a link where I can check how to write the proc compare statement?

https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/proc/n1xrf09nbjnxjzn1kdmbcajf1s1v.htm

https://documentation.sas.com/doc/en/pgmsascdc/9.4_3.5/proc/n0c1y14wyd3u7yn1dmfcpaejllsn.htm

Re: Proc compare to compare the data sets

ballardw — Thu, 21 Jul 2022 15:53:27 GMT

@Sandeep77 wrote:

Hi Experts,

I want to compare two huge data set with lot of rows and columns. With proc compare can I compare the entire data set or I have to select few columns and then compare? Also, is there a link where I can check how to write the proc compare statement?

If you have any doubt about the order of data you may need to consider sorting by variables at sufficient level to identify unique records. If the data is identical other than sort order you will get many mismatches because record 1 is compared with record 1, 2 with 2, 3 with 3 by default.