|
Join input data used in transformation for final output data
%join (indata = input data set,
outdata = output data set,
outlabel = output label,
keys = key variables,
dropmis
= drop missing values,
type = join type,
output = output program);
| Where |
Is
Type... |
And
represents... |
| indata |
C |
Source data set to be converted. This
is specified with a two level dot notation: libname.dataname.
Multiple datasets will be separated by spaces. |
| outdata |
C |
Destination data set.
This is specified with a two level dot notation: libname.dataname. |
| outlabel |
C (optional) |
Output dataset label name. |
| keys |
C |
Key variables separated by spaces. |
| dropmis |
C (optional) |
After the data is appended or merged, there
may be many missing rows that can be dropped. This can contain a
variable name or a list of variables separated by spaces. If all
the values of the specified fields are missing, the row will be
deleted. This is with the exception that the current row is the
only row for the key field(s).
|
| type |
C (optional) |
Data joining type. This includes:
|
| output |
C (optional) |
Name of the output SAS program. This
contains the name of the SAS program that contains the logic of the
program. If this were left blank, the program will be
saved to the current directory with the name join.sas.
|
Details
The data join utility will "join" data sets
specified into a specified destination data. This is commonly used
for data transformed by various transformation models including custom
transformations. The default append join type feature will append
data in the following way.
| INDATA1 |
|
|
INDATA2 |
|
|
var1 |
var2 |
|
var1 |
var3 |
|
value1 |
value2 |
|
value3 |
value4 |
|
value3 |
value4 |
|
value4 |
value5 |
|
|
|
|
|
|
|
OUTDATA |
|
|
|
|
var1 |
var2 |
var3 |
|
|
|
value1 |
value2 |
|
|
|
|
value3 |
value4 |
|
|
|
|
value3 |
|
value4 |
|
|
|
value4 |
|
value5 |
|
|
If the type is set to "merge" the result would look like:
|
INDATA1 |
|
|
INDATA2 |
|
|
var1 |
var2 |
|
var1 |
var3 |
|
value1 |
value2 |
|
value3 |
value4 |
|
value3 |
value4 |
|
value4 |
value5 |
|
|
|
|
|
|
|
OUTDATA |
|
|
|
|
var1 |
var2 |
var3 |
|
|
|
value1 |
value2 |
|
|
|
|
value3 |
value4 |
value4 |
|
|
|
value4 |
|
value5 |
|
|
Note that in a one to many relationships, the values will be repeated.
The utility will create a warning message in a many to many join but the
data will be merged with duplicates retained. In addition to creating a
base code program which produces the join, an optional code will be
created to reverse the join to re-produce the original dataset before the
merge. One extra reverse index dataset will be created to track
which dataset the reverse join will be located. Example %join (indata = inlib.adverse inlib.ae2,
outdata = outlib.ae ,
key = usubjid,
output = c:\temp\test.sas); |
