>On Sun 2002-08-18 at 18:53:46 +0200, arjones@stripped wrote:
>> To Whom it May Concern,
>> I sincerely doubt it's relevant, but i'm running MySQL 3.23.52 on
>> Linux 2.4.18 (RedHat build 5). Client and server are on the
>> and communicate via a Unix domain socket.
>> Perhaps it is a conscious decision, but it would seem an odd one:
>> When using LOAD DATA [LOCAL] INFILE, if the table has a column that is
>> BIGINT UNSIGNED NOT NULL AUTO_INCREMENT PRIMARY KEY and the infile uses \N
>\N means NULL with LOAD DATA. So you try to insert NULL values into a
>NOT NULL column. I find warnings about this quite reasonable.
I do too, but on the other hand, multi-row INSERT also reports info
including warnings and errors, and inserting NULL into an auto-increment
column isn't counted as a warning (I just tried it). I guess that might
be considered inconsistent...
>> for that column, a warning is generated for every line inserted. Replacing
>> \N with numbers in the infile gets rid of all warnings on import. Seems
>> to be a matter of adding one little check to Field_longlong::store() in
>> field.cc, but i could be wrong. I strongly suspect that it is
>> independant of integer size (e.g TINYINT, INT, BIGINT probably all
>> behave the same way).
>> Please send all replies to my personal address as well, as i am not
>> subscribed to the list.