Re: encoding error in new Postgres version

Hello,

nobody did answer so far it seems ...

> In 8.2 they had no problem. In 8.3, however they now get the following
> error
>
> On several tables:

> pg_restore: [archiver (db)] COPY failed: ERROR: character 0xe28899
> of encoding "UTF8" has no equivalent in "WIN1252"
> CONTEXT: COPY emails, line 3923

8.3 is more strict about encodings and locales. What encoding do source
(UTF8?) and target (WIN1252?) data base use? What are the locales (use
binary program pg_controldata on the cluster directories)?

Dump the problematic tables directly in plain text format and go to the
line which causes the error (e.g. line 3923 of data) and check for
characters not defined in WIN1252. If you use an ASCII editor you'll
most likely see a character that is used for 'undefined' characters - a
hollow square, a question mark etc.
http://en.wikipedia.org/wiki/Windows-1252
In your example it seems to be this character:
http://decodeunicode.org/en/u+2219/properties
Clean in it up in the source table if you've identified any problematic
character.

Good luck,

Peter