Different results from similar non-legal uses of OR operator

[Ken Chui]

Hello all,

I got a question about why these incorrect uses of the OR operator would end up with behaviors. I think because this is not a proper use, so I was not able to find any clue in the documentation. Here are the data and the incorrect codes:

clear
set obs 20
set seed 776
gen x = runiform() >= 0.5

Question: Why does the first code below show only cases where x = 1, but the second code prints everything?

. list if x == 1|0

     +---+
     | x |
     |---|
  1. | 1 |
  2. | 1 |
  3. | 1 |
  9. | 1 |
 13. | 1 |
     |---|
 17. | 1 |
 18. | 1 |
     +---+

.
. list if x == 0|1

     +---+
     | x |
     |---|
  1. | 1 |
  2. | 1 |
  3. | 1 |
  4. | 0 |
  5. | 0 |
     |---|
  6. | 0 |
  7. | 0 |
  8. | 0 |
  9. | 1 |
 10. | 0 |
     |---|
 11. | 0 |
 12. | 0 |
 13. | 1 |
 14. | 0 |
 15. | 0 |
     |---|
 16. | 0 |
 17. | 1 |
 18. | 1 |
 19. | 0 |
 20. | 0 |
     +---+

I fully understand that this is not correct, but am curious why the different outcomes.

Thanks, and I look forward to your thoughts.

[Nick Cox]

 
 list if x == 0|1

is parsed as

 
 list if (x == 0)|1

and while x == 0 is true contingently 1 is true (because non-zero) universally.

The fallacy is that the first statement is not equivalent to

list if x == 0 | x == 1

For more on true and false in Stata see e.g. https://www.stata.com/support/faqs/d...rue-and-false/ or https://www.stata-journal.com/articl...article=dm0087

[Andrew Musau]

The correct usage should be

list if x == 1|x==0

Your current commands state "list if x=0" or "if TRUE" and "list if x= 1" or "if FALSE". All positive numbers and missing evaluate to "TRUE", 0 evaluates to "FALSE".

clear
set obs 20
set seed 776
gen x = runiform() >= 0.5

list if 1
list if .
list if 0

Res.:

. list if 1

     +---+
     | x |
     |---|
  1. | 1 |
  2. | 1 |
  3. | 1 |
  4. | 0 |
  5. | 0 |
     |---|
  6. | 0 |
  7. | 0 |
  8. | 0 |
  9. | 1 |
 10. | 0 |
     |---|
 11. | 0 |
 12. | 0 |
 13. | 1 |
 14. | 0 |
 15. | 0 |
     |---|
 16. | 0 |
 17. | 1 |
 18. | 1 |
 19. | 0 |
 20. | 0 |
     +---+

.
. list if .

     +---+
     | x |
     |---|
  1. | 1 |
  2. | 1 |
  3. | 1 |
  4. | 0 |
  5. | 0 |
     |---|
  6. | 0 |
  7. | 0 |
  8. | 0 |
  9. | 1 |
 10. | 0 |
     |---|
 11. | 0 |
 12. | 0 |
 13. | 1 |
 14. | 0 |
 15. | 0 |
     |---|
 16. | 0 |
 17. | 1 |
 18. | 1 |
 19. | 0 |
 20. | 0 |
     +---+

.
. list if 0

.

Crossed with #2

[Nick Cox]

Let's be clear: In Stata illegal uses are flagged as syntax errors. Code that works without syntax error but does not do what you want is incorrect from your point of view but not illegal.

[Ken Chui]

Let's be clear: In Stata illegal uses are flagged as syntax errors. Code that works without syntax error but does not do what you want is incorrect from your point of view but not illegal.

Thank you, Nick and Andrew!

I think I got it now.

And Nick, thanks for the advice on the accurate wording regarding "legal".

[Nick Cox]

As a further note, it might be guessed from the fact that

list if x == 0 | 1

is parsed as

 list if (x == 0) | 1

that we just need parentheses to insist on what is wanted, namely

list if x == (0 | 1)

but that is not going to work either. because

0 | 1 

is just evaluated as 1. There are alternatives to

x == 0 | x == 1

such as

inlist(x, 0, 1)

or

inrange(x, 0, 1)

-- where the last is correct if and only if there are no values between 0 and 1.