mla
mla
Hrm. I see we already set the encoding of STDOUT, so I doubt that change will do anything. What value of LANG do you have in your shell?
When you view that row containing "NEW BLINDS: Supply & Delivery..." in psql, what does it look like? Do you see those messed up characters or something else?
k, thanks. So in the source db, what does that row look like? Does it show up as "Blinds Online�Home Brand\* *" there too?
SELECT company FROM ja_customers WHERE company LIKE 'NEW BLINDS%' ?
ok, good. So it seems like there's some encoding issue going on. We're sure the encoding of the source DB is UTF8? "psql -l" shows what for the source db?
And that string is exactly 255 chars. So the theory about the multibyte being interpreted as individual chars still makes sense then.
And if you grep the pg_sample output file for that string, you see the screwed up version, yes? grep "NEW BLINDS" all_data_3.sql
What perl is being used? /usr/bin/env perl -v
You might try a more recent perl. I'm using 5.18 for this testing, which is itself fairly old now. You can use the "plenv" tool to install local versions of...
What version of DBD::Pg do you have installed? perl -MDBD::Pg -le 'print DBD::Pg->VERSION'