Hello again,
I have some additional information that points to an
encoding problem. I pulled the COPY sql from the dump
file for the table that didn't get inserted properly
and tried to run just that section (with psql -f). I
got this error:
psql -f bibent_data.sql HRS
psql:bibent_data.sql:1: ERROR: copy: line 15, Invalid
UNICODE character sequence found (0xc876)
psql:bibent_data.sql:1: lost synchronization with
server, resetting connection
I'm running on FreeBSD 4.5. Is there a way I need to
recode the file before running it?
Thanks,
Paul Warner
Paul Warner wrote:
> Hello,
>
> I've looked for this in the books and archives and
> haven't been able to find an answer.
>
> I did a pg_dump on an older postgres database (I
> believe it's 7.0.something), and restored to 7.2.1
> with multibyte capability. The old db is not
> multibyte. What's happened is that there are a
> significant number of rows in the dump file that are
> not getting loaded into the new database. The tables
> are getting there, but not the data.
>
> I notice that there are triggers added in 7.2 that
> don't appear in the earlier version (as in: Triggers:
> RI_ConstraintTrigger_1663276). I don't know if this
> is related, but it is true that all the 952 messages
> related to a certain topic didn't make it into one
> table.
>
> This strikes me as a version issue related to
> integrity constraints, but could it be related to the
> change to multibyte?
>
> I'm working on FreeBSD 4.5, with aolserver and openacs
> 3.2.4.
>
> Thanks very much for your help!
>
> Paul Warner
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org