SAS Chapter 18: Modifying SAS Data Sets And Tracking Changes

UNIQUE

CHECK

PRIMARY KEY

NOT NULL

proc datasets nolist; audit quarter1; initiate; quit;

proc datasets lib=reports nolist; audit initiate reports.quarter1; quit;

proc datasets lib=reports nolist; initiate audit quarter1; quit;

proc datasets lib=reports nolist; audit quarter1; initiate; quit;

data records.snowfall; modify records.snowfall records.overnight key=accumAmt; run;

data records.snowfall; modify records.overnight records.snowfall; by accumAmt; run;

data records.snowfall; modify records.snowfall records.overnight; by accumAmt; run;

data records.snowfall; modify records.snowfall records.overnight; update accumAmt; run;

To determine whether the index specified on the KEY= option is a valid index

To determine the number of observations that were updated in the master data set

To determine the status of the I/O operation

To determine the number of observations that could not be updated in the master data set

proc datasets nolist; modify age; ic create val_age=check(where=(age>=18)); quit;

proc datasets nolist; modify Survey; ic create val_age=check(age>=18); quit;

proc datasets nolist; modify survey; integrity constraint val_age=check(where=(age>=18)); quit;

proc datasets nolist; modify survey; ic create val_age=check(where=(age>=18)); quit;

MODIFY creates a second copy of the data while variables in the data are being matched with a WHERE clause and then deletes the second copy.

You cannot modify the descriptor portion of the data set using the MODIFY statement.

You can use the MODIFY statement to change the name of a variable.

If the system terminates abnormally while a DATA step that is using the WHERE statement is processing, SAS automatically saves a copy of the unaltered data set.

They create historical versions of data so that a copy of the data set is saved each time the data is replaced.

They record information about changes to observations in a data set each time the data set is replaced.

They record information about changes to observations in a data set each time the data is modified in place.

The audit trail file has the same name as the SAS data file it is monitoring, but has #AUDIT at the end of the data set name.

proc datasets lib=sasuser nolist; modify Amounts (genmax=6); quit;

proc datasets lib=sasuser nolist; modify Amounts (genmax=5); quit;

proc datasets lib=sasuser nolist; modify Amounts (gennum=6); quit;

proc datasets lib=sasuser nolist; modify Amounts (gennum=5); quit;

SAS locates the variables to update using the index specified in the KEY= option and then automatically overlays nonmissing transaction values as it does when you use the MODIFY/BY statements.

When you use the KEY= option, you must explicitly state the update that you want to make. SAS does not automatically overlay nonmissing transaction values.

The KEY= option is used to specify a variable to match for updating observations.

The index named in the KEY= option must be a simple index.

proc datasets lib=sasuser nolist; delete amounts (gennum=ALL); quit;

proc datasets lib=sasuser nolist; delete amounts (gennum=HIST); quit;

proc datasets lib=sasuser nolist; delete amounts (gennum=0); quit;

proc datasets lib=sasuser nolist; delete amounts; quit;