Delete last. reporting_date

[Babloo]

I've a data as follows as I want to delete the last record of the reporting date 10/09/2018.

Input data:

Reporting_date	Error_id
10/09/2018	12
10/09/2018	24
11/09/2018	11
12/09/2018	13
13/09/2018	16

Program:

proc sql;
create table test as select * from dex.FUNCTIONAL_ERRORS;
quit;

proc sort data=test;
by reporting_date error_id;
run;

data temp;
set test;
by reporting_date error_id;
if last.reporting_date=10/09/2018 then delete;
/*else new_reporting_date;*/
run;

Extract of log:

33         
34         data temp;
35         set test;
36         by reporting_date error_id;
37         if last.reporting_date=10/09/2018 then delete;
38         /*else new_reporting_date;*/
39         run;

NOTE: There were 5 observations read from the data set WORK.NEW_TEST.
NOTE: The data set WORK.TEMP has 5 observations and 2 variables.
NOTE: DATA statement used (Total process time):
      real time           0.01 seconds
      cpu time            0.01 seconds

May I know why the last record of 10/09/2018 is not getting deleted? Reporting _date is numeric variable with format ddmmyy10.

Desired Output:

Reporting_date	Error_id
10/09/2018	12
11/09/2018	11
12/09/2018	13
13/09/2018	16

[gamotte]

Hello last.var is an automatic variable created by the "by var;" instruction.

last.var=1 if the observation is the last one for the current value of var, 0 otherwise.

It makes no sense to compare it to a date.

if last.reporting_date=10/09/2018

should be  (assuming reporting_date has a date format)

if last.reporting_date and reporting_date='10Sep2018'd

[Babloo]

It is not solving the issue.

6         data temp;
27         set test;
28         by reporting_date error_id;
29         if reporting_date and last.reporting_date='10sep2018'd then delete;
30         /*else new_reporting_date;*/
31         run;

NOTE: There were 7851 observations read from the data set WORK.TEST.
NOTE: The data set WORK.TEMP has 7851 observations and 12 variables.

[gamotte]

You are still comparing last.reporting_date, which is 0/1, to a date.

Try :

if last.reporting_date and reporting_date='10sep2018'd then delete;

[Astounding]

It seems you are describing only a small part of the problem.

• What should happen if there are three observations for 10/09/2018 instead of two?
• What should happen if some other date has three observations?

Given that you sort your data, here are a few ways to subset that you might want to use in the final DATA step:

if first.reporting_date;

Keeps just one observation per reporting date, whichever one is first.

if last.reporting_date and 0=first.reporting_date then delete;

Deletes the last observation per reporting date, as long as there is more than one.

The right solution depends on accurately describing the problem.  You may have done that, but I suspect you did it for just a small selection of the data.