*Copyright @ www.mycsg.in;

Pre-requisite

This lesson covers appending datasets using PROC APPEND
For an introduction to appending concepts, common scenarios, and appending using the SET statement, refer to lesson SAS_APPENDING_L101

What is PROC APPEND?

PROC APPEND is a SAS procedure specifically designed to append observations from one dataset to another
It adds observations from the DATA dataset to the end of the BASE dataset
Unlike the SET statement, PROC APPEND does not read the BASE dataset - it only reads the DATA dataset
This makes PROC APPEND significantly more efficient than the SET statement when the BASE dataset is large
A common use case is appending a new batch of records to a large accumulating dataset

Syntax of PROC APPEND

The basic syntax for PROC APPEND requires two arguments: BASE= specifies the target dataset that will receive the appended observations, and DATA= specifies the source dataset whose observations will be added
Optional arguments such as FORCE can follow the DATA= option when needed

/*
proc append base=base-dataset data=data-dataset <options>;
run;
*/

BASE= specifies the dataset to which observations are to be appended (the target dataset)
DATA= specifies the dataset whose observations are to be added to the BASE dataset
The BASE dataset is modified directly - no separate output dataset is created

Key difference between SET statement and PROC APPEND

SET statement reads ALL input datasets to create a new output dataset
PROC APPEND reads only the DATA dataset and appends it to the BASE dataset without reading the BASE dataset
When the BASE dataset has millions of records and DATA has a small batch of new records, PROC APPEND is much faster
With SET statement, as the BASE grows larger, the time to append also grows - with PROC APPEND it does not

Create example datasets

Recreating the same example datasets used in SAS_APPENDING_L101 for consistency

data males females;
   set sashelp.class;
   if sex="M" then output males;
   if sex="F" then output females;
run;
 
data sedan sports othercars;
   set sashelp.cars;
   if type="Sedan" then output sedan;
   else if type="Sports" then output sports;
   else output othercars;
run;

Appending datasets with same attributes using PROC APPEND

When BASE and DATA datasets have the same variables with the same attributes, PROC APPEND works without any additional options
Observations from the DATA dataset are added to the end of the BASE dataset
The BASE dataset is updated in place - it will contain observations from both the original BASE and the DATA dataset after the procedure runs

Append females dataset to males dataset using PROC APPEND

Before PROC APPEND, males has 10 observations and females has 9 observations
After PROC APPEND, males will contain 19 observations - original 10 male records followed by 9 female records
The females dataset remains unchanged

proc append base=males data=females;
run;

Append sports dataset to sedan dataset using PROC APPEND

Before PROC APPEND, sedan has 262 observations and sports has 120 observations
After PROC APPEND, sedan will contain 382 observations

proc append base=sedan data=sports;
run;

PROC APPEND with FORCE option

By default, PROC APPEND requires the BASE and DATA datasets to have the same variables with compatible attributes
If the DATA dataset has variables not present in the BASE dataset, PROC APPEND produces an error and stops
If a variable has a longer length in DATA than in BASE, PROC APPEND produces an error and stops
The FORCE option instructs PROC APPEND to proceed despite these differences
When FORCE is used and DATA has extra variables not in BASE, those variables are dropped during the append
When FORCE is used and a variable is longer in DATA than in BASE, values may be truncated - always check the log for warnings

Appending datasets with different variables using FORCE option

Creating two datasets where the DATA dataset has an extra variable not present in the BASE dataset

data class_with_bmi;
   set sashelp.class;
   bmi = (weight / (height ** 2)) * 703;
run;
 
data class_no_bmi;
   set sashelp.class(obs=5);
run;

class_with_bmi contains all class variables plus BMI
class_no_bmi contains only the standard class variables
Appending class_with_bmi to class_no_bmi without FORCE would produce an error because BMI is not in the BASE dataset
With FORCE, the append proceeds and the BMI variable is dropped from the appended records

proc append base=class_no_bmi data=class_with_bmi force;
run;

Appending datasets where DATA has a longer variable length than BASE

Creating two datasets where the same variable has different lengths

data name_short;
   set sashelp.class(obs=5);
   keep name age;
run;
 
data name_long;
   length name $20;
   set sashelp.class;
   name = catx(" ", name, "Johnson");
   keep name age;
run;

name_short has NAME with default length of 8 characters
name_long has NAME with length of 20 characters
Without FORCE, PROC APPEND would error because the DATA variable NAME is longer than in BASE
With FORCE, the append proceeds but NAME values from name_long are truncated to 8 characters
Always check the SAS log for truncation warnings when using FORCE

proc append base=name_short data=name_long force;
run;

*Copyright @ www.mycsg.in;

Pre-requisite

What is PROC APPEND?

Syntax of PROC APPEND

SAS Log

Key difference between SET statement and PROC APPEND

Create example datasets

SAS Log

Appending datasets with same attributes using PROC APPEND

Append females dataset to males dataset using PROC APPEND

SAS Log

Dataset View

Append sports dataset to sedan dataset using PROC APPEND

SAS Log

Dataset View

PROC APPEND with FORCE option

Appending datasets with different variables using FORCE option

SAS Log

SAS Log

Dataset View

Appending datasets where DATA has a longer variable length than BASE

SAS Log

SAS Log

Dataset View