Thread: BUG #8141: multi-column check expression evaluating to NULL

The following bug has been logged on the website:

Bug reference:      8141
Logged by:          Andras Vaczi
Email address:      andras.vaczi@zalando.de
PostgreSQL version: 9.1.9
Operating system:   linux/Ubuntu 12.10
Description:        =


Consider the following table with a CHECK constraint:

CREATE TABLE check_test
(
  id integer NOT NULL,
  col integer,
  CONSTRAINT unique_with_null_check1 CHECK (col >=3D 1 AND id < 20)
);

This INSERT statement succeeds:

INSERT INTO check_test (id, col) VALUES (1, NULL);

While, col being NULL, the whole CHECK condition evaluates to NULL - this is
covered in the documentation.

But this is refused:
INSERT INTO check_test (id, col) VALUES (21, NULL);

ERROR:  new row for relation "check_test" violates check constraint
"unique_with_null_check1"

I think this behaviour should be either also mentioned in the docs or
cosidered a bug.

andras.vaczi@zalando.de writes:
> Consider the following table with a CHECK constraint:

> CREATE TABLE check_test
> (
>   id integer NOT NULL,
>   col integer,
>   CONSTRAINT unique_with_null_check1 CHECK (col >= 1 AND id < 20)
> );

> This INSERT statement succeeds:

> INSERT INTO check_test (id, col) VALUES (1, NULL);

> While, col being NULL, the whole CHECK condition evaluates to NULL - this is
> covered in the documentation.

> But this is refused:
> INSERT INTO check_test (id, col) VALUES (21, NULL);

> ERROR:  new row for relation "check_test" violates check constraint
> "unique_with_null_check1"

> I think this behaviour should be either also mentioned in the docs or
> cosidered a bug.

I see no bug here.  In the first case, the "col >= 1" condition yields
NULL while "id < 20" yields TRUE, so you have NULL AND TRUE which is
NULL, which is considered a "pass" for a CHECK condition per spec.
In the second case, "col >= 1" is still NULL, but "id < 20" is FALSE,
so you have NULL AND FALSE which is FALSE (*not* NULL), and so failure
is per spec.

Yes, the behavior of AND/OR with NULLs is documented.
http://www.postgresql.org/docs/9.1/static/functions-logical.html

            regards, tom lane

On 09/05/13 01:07, andras.vaczi@zalando.de wrote:
> The following bug has been logged on the website:
>
> Bug reference:      8141
> Logged by:          Andras Vaczi
> Email address:      andras.vaczi@zalando.de
> PostgreSQL version: 9.1.9
> Operating system:   linux/Ubuntu 12.10
> Description:
>
> Consider the following table with a CHECK constraint:
>
> CREATE TABLE check_test
> (
>    id integer NOT NULL,
>    col integer,
>    CONSTRAINT unique_with_null_check1 CHECK (col >= 1 AND id < 20)
> );
>
[...]

if 'id' is the primary key, then the table should be defined as:

CREATE TABLE check_test
(
   id  integer PRIMARY KEY,
   col integer NOT NULL,
   CONSTRAINT unique_with_null_check1 CHECK (col >= 1 AND id < 20)
);

Because:

 1.   'PRIMARY KEY ' implies 'NOT NULL' and also 'UNIQUE', and so an
    index is created for 'id'
    (index is required to enforce uniqueness)

 2. you are testing that 'col' has a certain type of numeric value, so
    this logically implies that it is 'NON NULL', but you actually have
    to tell progress explicitly tat you want it to be 'NON NULL'



Cheers,
Gavin