BookmarkSubscribeRSS Feed
heyyou1
Fluorite | Level 6

Hello, first off thanks for your help. The SAS community has been very helpful so far.

 

I have data that looks like this

County | Region | FIPS | Race | Applied Count
"Cool County" | 1 | 10101 | Asian | 5
"Cool County" | 1 | 10101 | Black or African American| 50
"Cool County" | 1 | 10101 | Other Race| 20
"Cool County" | 1 | 10101 | Unknown | 8
"Cool County" | 1 | 10101 | White | 30

 

I then have census population data that looks like this:

County | Region | FIPS | Race | Population
"Cool County" | 1 | 10101 | Asian | 6000
"Cool County" | 1 | 10101 | Black or African American| 35000
"Cool County" | 1 | 10101 | Other Race/ Unknown | 18000
"Cool County" | 1 | 10101 | White | 200000

 

I would like to combine the tables into something like this, so I can understand what percentage of the population applied.

County | Region | FIPS | Race | Applied Count | Population

"Cool County" | 1 | 10101 | Asian | 5 | 6000
"Cool County" | 1 | 10101 | Black or African American| 50 | 35000
"Cool County" | 1 | 10101 | Other Race/ Unknown | 28 | 18000
"Cool County" | 1 | 10101 | White | 30 | 200000

 

How do I merge the "unknown" and "other race" rows in the Application Count table to better merge with the census data? 

 

All other rows line up and have a match and will have a match in the future. It's just the oddity of the "unknown" and "other race" rows.

3 REPLIES 3
PaigeMiller
Diamond | Level 26

You could use a text editor to globally change 

 

Other Race/ Unknown

into 

Other Race

Or you could do the same thing in a SAS data step

--
Paige Miller
heyyou1
Fluorite | Level 6
How would I do that with a data step while summing "Applied count" for both of them?
Reeza
Super User

For your first data set:

 

data recoded;
set table1;
if race  in ( 'Other Race', Unknown') then race = 'Other Race/Unknown';
run;

proc means data=recoded noprint SUM;
by county region fips race;
var applied_count;
ods output summary = summarized;
run;

Then do a standard match merge process of your choice.

 


@heyyou1 wrote:

Hello, first off thanks for your help. The SAS community has been very helpful so far.

 

I have data that looks like this

County | Region | FIPS | Race | Applied Count
"Cool County" | 1 | 10101 | Asian | 5
"Cool County" | 1 | 10101 | Black or African American| 50
"Cool County" | 1 | 10101 | Other Race| 20
"Cool County" | 1 | 10101 | Unknown | 8
"Cool County" | 1 | 10101 | White | 30

 

I then have census population data that looks like this:

County | Region | FIPS | Race | Population
"Cool County" | 1 | 10101 | Asian | 6000
"Cool County" | 1 | 10101 | Black or African American| 35000
"Cool County" | 1 | 10101 | Other Race/ Unknown | 18000
"Cool County" | 1 | 10101 | White | 200000

 

I would like to combine the tables into something like this, so I can understand what percentage of the population applied.

County | Region | FIPS | Race | Applied Count | Population

"Cool County" | 1 | 10101 | Asian | 5 | 6000
"Cool County" | 1 | 10101 | Black or African American| 50 | 35000
"Cool County" | 1 | 10101 | Other Race/ Unknown | 28 | 18000
"Cool County" | 1 | 10101 | White | 30 | 200000

 

How do I merge the "unknown" and "other race" rows in the Application Count table to better merge with the census data? 

 

All other rows line up and have a match and will have a match in the future. It's just the oddity of the "unknown" and "other race" rows.


 

SAS Innovate 2025: Save the Date

 SAS Innovate 2025 is scheduled for May 6-9 in Orlando, FL. Sign up to be first to learn about the agenda and registration!

Save the date!

Mastering the WHERE Clause in PROC SQL

SAS' Charu Shankar shares her PROC SQL expertise by showing you how to master the WHERE clause using real winter weather data.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 3 replies
  • 597 views
  • 0 likes
  • 3 in conversation