Hi Jasen,
Jasen Betts wrote:
>
> why not just copy the whole lot into a temp table (like the target
> table but with text fields instead of fixed-width varchar)
> and then do a select on that to find the over-length lines and another
> select to insert the good data into the real target table.
>
Thanks, we did that. In fact the postgresql db schema is generated from
a xml file, so we changed fields where this applies to now be text and
made them bigger in postgresql schema. Now the nightly mirroring runs
through. without any field 'over-runs' stopping it.
Regards,
--
Urs Rau
