John R Pierce wrote:
> On 7/4/2014 2:12 AM, sunpeng wrote:
>> Thank you, friend, I use --hex-blob :
>> mysqldump -v -nt --complete-insert=TRUE --compatible=postgresql
>> --default-character-set=utf8 --skip-add-locks --compact
>> --no-create-info --skip-quote-names --hex-blob -uroot -p test
>> videorecresult >dbdata.sql
>> to dump mysql data.
>> And replace blob data "0x...." into "E'\\xx....'" to load data into
>> postgresql.
> regardless of all that, 0xFF is not a valid UTF8 character code. perhaps
> you should store the data in a postgres BYTEA, or at least use character
> encoding SQLASCII (which is to say, no encoding, bytes in == bytes out.)
Exactly.
According to mysqldump's man page, the affected field must be
BINARY, VARBINARY, the BLOB types or BIT.
For these PostgreSQL's "bytea" would definitely be the correct data type,
and there wouldn't be any encoding problems with that.
Yours,
Laurenz Albe