Selecting Records Based on Appearances

[BCNAV]

This is a bit of a weird request....I have some data on aircraft flights that I need to break out depending on how many times they appear in a group. The group is identified as follows:

daily_grouping
104042452
104042452
104042452
104042452
104042452
104042452
104042453
104042454
104042455
104042455
104042456
104042456
104042456

There are other variables as well, but the daily_grouping is the one to key in on. Here is what I would like:

1. If a daily_grouping number appears only once, it should be removed.

2. If a daily_grouping number appears more than once, then the first instance should be removed and n-1 should remain.

This means that in the example above the result would be:

daily_grouping
104042452
104042452
104042452
104042452
104042452
104042455
104042456
104042456

Any help would be appreciated! Not sure if multiple steps are needed.

[ballardw · Accepted solution]

data have;
input daily_grouping :$10.;
datalines;
104042452
104042452
104042452
104042452
104042452
104042452
104042453
104042454
104042455
104042455
104042456
104042456
104042456
;

data want;
   set have;
   by daily_grouping;
   if not first.daily_grouping;
run;

[andreas_lds]

Untested code:

data want;
set have;
by daily_grouping;
if not first.daily_grouping;
run;

[BCNAV]

thanks all!

[Ksharp]

data have;
input daily_grouping :$10.;
datalines;
104042452
104042452
104042452
104042452
104042452
104042452
104042453
104042454
104042455
104042455
104042456
104042456
104042456
;

proc sort data=have dupout=want nodupkey;
by daily_grouping ;
run;