If I group the data I have by name and id, I want to add a new row per group if the first code = last code, the new row need to have the same name and id, and the value of the 'amount' column is N/A, the values for the rest columns can just be the same as one of the rows in the group (can also just be empty, doesn't really matter).
data have;
input
name : $1.
id : 8.
code : $8.
amount : 8.
;
datalines;
A 10 12345678 100
A 10 09876543 30
B 9 88997700 20
B 9 88997700 -20
B 9 88997700 40
C 28 11223344 30
C 28 11223344 40
C 28 67676767 50
D 10 78654890 40
D 10 78654890 50
;
data want;
input
name : $1.
id : 8.
code : $8.
amount : 8.
;
datalines;
A 10 12345678 100
A 10 09876543 30
B 9 88997700 20
B 9 88997700 -20
B 9 88997700 40
B 9 88997700 N/A
C 28 11223344 30
C 28 11223344 40
C 28 67676767 50
D 10 78654890 40
D 10 78654890 50
D 10 78654890 N/A
;
If the "amount" is a numeric variable as shown in your data step you cannot have a value of "N/A". You can have a missing value. If you want to see a displayed value of "N/A" then you can create format that will do such but the value can't be text.
This makes a set that looks like that when the format MYNA is available. You would have to make sure the format code is run in any session you want to see N/A, or learn about format management to create and use formats stored in permanent libraries.
data have; input name : $1. id : 8. code : $8. amount : 8. ; datalines; A 10 12345678 100 A 10 09876543 30 B 9 88997700 20 B 9 88997700 -20 B 9 88997700 40 C 28 11223344 30 C 28 11223344 40 C 28 67676767 50 D 10 78654890 40 D 10 78654890 50 ; proc format; value myna .N='N/A' ; run; data want; set have; by name id; length firstcode $ 8; retain firstcode; if first.id then firstcode=code; if last.id and code=firstcode then do; /* write the current values to the data set*/ output; /* assign missing to amount*/ amount = .N; output; end; else output; format amount myna. ; drop firstcode; run;
SAS has 27 "special" missing values from .A to .Z plus ._ in addition to the default . missing. You can assign formats to display different meanings for each of them for a single variable. The special missing values will appear as A to Z or _ if a format is not assigned to display them differently.
If the "amount" is a numeric variable as shown in your data step you cannot have a value of "N/A". You can have a missing value. If you want to see a displayed value of "N/A" then you can create format that will do such but the value can't be text.
This makes a set that looks like that when the format MYNA is available. You would have to make sure the format code is run in any session you want to see N/A, or learn about format management to create and use formats stored in permanent libraries.
data have; input name : $1. id : 8. code : $8. amount : 8. ; datalines; A 10 12345678 100 A 10 09876543 30 B 9 88997700 20 B 9 88997700 -20 B 9 88997700 40 C 28 11223344 30 C 28 11223344 40 C 28 67676767 50 D 10 78654890 40 D 10 78654890 50 ; proc format; value myna .N='N/A' ; run; data want; set have; by name id; length firstcode $ 8; retain firstcode; if first.id then firstcode=code; if last.id and code=firstcode then do; /* write the current values to the data set*/ output; /* assign missing to amount*/ amount = .N; output; end; else output; format amount myna. ; drop firstcode; run;
SAS has 27 "special" missing values from .A to .Z plus ._ in addition to the default . missing. You can assign formats to display different meanings for each of them for a single variable. The special missing values will appear as A to Z or _ if a format is not assigned to display them differently.
Learn how use the CAT functions in SAS to join values from multiple variables into a single value.
Find more tutorials on the SAS Users YouTube channel.
Ready to level-up your skills? Choose your own adventure.