Delete duplicates including the original one

[Dhiroj Prasad Koirala]

Hi,
I have been trying to delete observations, including the original one, in stata. I could not find any code. The 'drop duplicates var' deletes the duplicate observations if any, but keeps the first observation. BUT, I want to remove all observations that have duplicates, including the first one. Please help.

[Rich Goldstein]

without more, it is hard to be concrete but here is a strategy - use -duplicates tag- to tag one of the duplicates and then use -egen- with the max option to spread that tag to all relevant observations and then drop those so tagged (you can then drop the tag variable and the variable created by -egen-)

[Nick Cox]

In #1 drop duplicates should be duplicates drop

duplicates isn't designed for what you want, even as a special or limiting case: but what you want is easy. For the variables concerned you write

bysort varlist : drop if _N > 1

Here's a prototype, but watch out. This can be terribly destructive.

. clear

. set obs 10
number of observations (_N) was 0, now 10

. gen foo = word("a b c d e", _n) in 1/5
(5 missing values generated)

. replace foo = "f" in 6/L
(5 real changes made)

.
. list

     +-----+
     | foo |
     |-----|
  1. |   a |
  2. |   b |
  3. |   c |
  4. |   d |
  5. |   e |
     |-----|
  6. |   f |
  7. |   f |
  8. |   f |
  9. |   f |
 10. |   f |
     +-----+

.
. bysort foo : drop if _N > 1
(5 observations deleted)

.
. list

     +-----+
     | foo |
     |-----|
  1. |   a |
  2. |   b |
  3. |   c |
  4. |   d |
  5. |   e |
     +-----+