I am trying to migrate a modest sized database (the unload file is just under a gigabyte) which is moderately complex, using schemas and the like from 7.4 to 8.1 Alas, pg_dump appears to incompatable with my needs as it keeps introducing invalid UTF-8 characters. I have gone back and changed the offending data rows with SQL to make damned sure there are no illegal characters. I redump, compress the thing, scp it and try to reload. Same error every time: psql:glacier_load.sql:1017: ERROR: invalid UTF-8 byte sequence detected near byte 0x96 CONTEXT: COPY clients, line 1751, column client_name: "XYZ Media AB&E" psql:glacier_load.sql:3019: ERROR: current transaction is aborted, commands ignored until end of transaction block See that double space after "Media" ? That's actually a single space, a dash, and another space: "XYZ Media - AB&E", but every export turns it into: XYZ Media <96> AB&E Is there any way to do this ? I can't edit the resulting dump file by hand -- it's huge. And pg_dump on 7.4 does not know how to a table inside a schema -- it seems to be all or nothing. TIA, Greg Williamson DBA GlobeXplorer LLC
