On Wed, Mar 5, 2014 at 11:37 PM, Ian Lawrence Barwick <barwick@gmail.com> wrote:
> 2014-03-05 23:27 GMT+09:00 Andrew Dunstan <andrew@dunslane.net>:
>>
>> On 03/05/2014 09:11 AM, Michael Paquier wrote:
>>>
>>> After testing this feature, I noticed that FORCE_NULL and
>>> FORCE_NOT_NULL can both be specified with COPY on the same column.
>>> This does not seem correct. The attached patch adds some more error
>>> handling, and a regression test case for that.
>>>
>>
>>
>> Strictly they are not actually contradictory, since FORCE NULL relates to
>> quoted null strings and FORCE NOT NULL relates to unquoted null strings.
>> Arguably the docs are slightly loose on this point. Still, applying both
>> FORCE NULL and FORCE NOT NULL to the same column would be rather perverse,
>> since it would result in a quoted null string becoming null and an unquoted
>> null string becoming not null.
>
> Too frazzled to recall clearly right now, but I think that was the somewhat
> counterintuitive conclusion I originally came to.
In this case I may be an intuitive guy :), but OK I see your point. So
if we specify both this produces the exact opposite as the default,
default being an empty string inserted for a quoted empty string and
NULL inserted for a non-quoted empty string. So yes I'm fine with a
note on the docs about that, and some more regression tests.
Btw, if we allow this behavior in COPY, why doesn't file_fdw allow
both options to be allowed on the same column for a foreign table?
Current behavior of file_fdw seems rather inconsistent with COPY as it
stands now.
--
Michael
