Is there a way to extract overlapping observations from a dataset?

[sneha_dodle]

I need to find overlapping advevent value for the same patid and date.
This is the dataset i have:
Patid advevent date
1 vomiting 01jan09
1 vomiting 01jan09
1 vomiting 01jan09
2 delirium 06jan09
2 nausea 06jan09
2 delirium 06jan09
2 delirium 06jan09

My outpout should be

2 delirium 06jan09
2 nausea 06jan09

[PaigeMiller]

Please explain further. Why are those two records the only output? What is the logic that keeps those two lines and not the others?

[sneha_dodle]

My task to create a dataset that contains observations with different
advevent value for the same patid and visitdate.
Im creating an outliers dataset. A single patient cannot have two
different adverse events at the same visit date.

[Astounding]

I would recommend you go back and check on the task you were given.

It is certainly possible for a patient to have more than one adverse event on the same day.  It would be unusual to be allowed to report only one.

It is unusual for a patient to report the same adverse event multiple times on the same day.  If a patient reports 3 headaches on the same day, they would usually be treated as a single headache.

[PaigeMiller]

This does not address my earlier question: "Why are those two records the only output? What is the logic that keeps those two lines and not the others?"

Specifically, why are there no output records for Patid 1?

[sneha_dodle]

Ive given only a sample of my dataset. i have patid values that go upto 244 and 4 to 6 advevent values for each patid

[Shmuel]

Without entering, is the query logic or not, I assume that next code may give you the requested result:

proc sort data=have out=temp nodupkey;
  by patid date advevent;
run;

data want;
set temp;
   by patid date advevent;;
        if not (first.date and last.date);
run;

[s_lassen]

Here is an SQL solution:

proc sql;
  create table want as
  select distinct Patid, advevent,date from have
  group by Patid, date
  having count(distinct advevent)>1;
quit;