Chris Ruprecht <chris@cdrbill.com> writes:
> I can export the data out of that database into flat files just fine, but then I try to import the data to Postgres,
I'mgetting errors like this:
> ERROR: invalid byte sequence for encoding "SQL_ASCII": 0x00
> CONTEXT: COPY attachments, line 14: "58025 1 cl\Cert.r 10
M04P'15A415).($-H87)4:6UE+$-(05)!0U1%4BQ)3E!55"!I5&EM92!)3E1% M1T52'$585$523B!7..."
> I tried LATIN1, SQL_ASCII, UTF-8, nothing works. I even tried to make the data type 'bytea', no luck. I'd love to
havea "NO-CONVERSION" option on the copy command that just takes what ever bytes come along and doesn't try to
interpretthem.
That's hardly possible, considering you're expecting COPY to recognize
field and record boundaries. What you probably need to do here is declare
the column as bytea and then write some sort of preprocessing script that
converts the binary data into hex-encoded form (basically \x followed by
hex digits, if memory serves, but check the description of bytea in the
PG manual).
regards, tom lane
