Re: copy from .. How to get rid of encoding check for bytea coumns

Roman Golis wrote:
> I am trying to load data into a rather simple table:
>
> CREATE TABLE "public"."files" (
>   "id" SERIAL,
>   "idchar" CHAR(32) NOT NULL,
>   "content" BYTEA,
>   CONSTRAINT "files_pkey" PRIMARY KEY("id")
> ) WITHOUT OIDS;
>
> with this command:
>
> copy files (idchar, content) from '/data/1.dat' delimiter '|';
>
> The database encoding is UTF-8.
>
> Here is an example of the data file content:
>
>
0C2CCE6941194369B020000B616F1301|\xFF\xD8\xFF\xE0\x00\x10\x4A\x46\x49\x4
6\x00\x01
>
> And I get this error:
>
> ERROR:  invalid byte sequence for encoding "UTF8": 0xff
>
> The command
>
> set client_encoding = 'SQL_ASCII';
>
> Does not helps at all, the result is the same error message.
>
> How can I turn off that annoying codepage checking during COPY FROM ?
>
> Isn't it a bug ?

I assume that you are using PostgreSQL 9.1.

There are two mistakes:
1) You didn't tell COPY that you want CSV format.
2) Your bytea is not properly encoded

Fix your data file to look like this:
0C2CCE6941194369B020000B616F1301|\xFFD8FFE000104A4649460001

Then load it like this:
COPY files (idchar, content) FROM '/data/1.dat' (FORMAT 'csv', DELIMITER
'|');

The error message you see is because COPY thinks that the \xFF
is part of the "idchar" character column.

Yours,
Laurenz Albe