On Fri, Jan 23, 2009 at 02:18, David Goodenough
wrote:
>
> Is there a definative HOWTO that I can follow, if not does someone
> have a set of instructions that will work?
>
> If it matters I am running under Debian.
I did it once for a very large db (large for me was 5GB) and converted
it from EU
On 2009-01-22, David Goodenough wrote:
> I have a database which was created as LATIN1 (the machine has the
> wrong locales installed when I set up PG). It is running 8.3.
>
> So I found various places which said the way to do this was to do
> a pg_dumpall -f dump_file, get rid of the entire dat
> > cat the-source-dump.sql | iconv -t utf8 - > my-converted.sql
> >
> > Size should not matter in this case...
>
> Yeah it does. iconv buffers everything in memory, as I recall.
Just found an alternative - "uconv" command (part of ICU project):
http://www.icu-project.org/userguide/intro.html
Alan Hodgson wrote:
Yeah it does. iconv buffers everything in memory, as I recall.
However, you can "split" the file into manageable pieces, run each
through
iconv, and recombine afterwards.
Another way is to just use GNU recode for large files instead of iconv.
It's slower but do
On Thursday 22 January 2009, David Goodenough
wrote:
>
> You have not understood what I said. I ran iconv, and it changes the
> encoding of the data, but not the ENCODING= statements that are
> embedded in the datastream. Yes I can change those with sed, but
> I do not know what else I need to
> You have not understood what I said. I ran iconv, and it changes the
> encoding of the data, but not the ENCODING= statements that are
> embedded in the datastream. Yes I can change those with sed, but
> I do not know what else I need to change. There must be an easier
> way.
Oops, please a
On Thursday 22 January 2009, Vladimir Konrad wrote:
> > iconv does not change the database encodings embedded in the file
> > (and it is quite large).
>
> Have you read the manual?
>
>file A pathname of an input file. If no file operands are
>specified, or if a file operand is '-'
On Thursday 22 January 2009, Vladimir Konrad wrote:
> > iconv does not change the database encodings embedded in the file
> > (and it is quite large).
>
> Have you read the manual?
>
>file A pathname of an input file. If no file operands are
>specified, or if a file operand is '-
> iconv does not change the database encodings embedded in the file
> (and it is quite large).
Have you read the manual?
file A pathname of an input file. If no file operands are
specified, or if a file operand is '-', the standard input shall
be used.
cat the-source-dump
On Thursday 22 January 2009, Vladimir Konrad wrote:
> > Is there a definative HOWTO that I can follow, if not does someone
> > have a set of instructions that will work?
>
> What about running "iconv" command on the dumped .sql file and transform
> it to the utf8?
>
> Vlad
>
> PS: man iconv for man
> Is there a definative HOWTO that I can follow, if not does someone
> have a set of instructions that will work?
What about running "iconv" command on the dumped .sql file and transform
it to the utf8?
Vlad
PS: man iconv for manual
--
Sent via pgsql-general mailing list (pgsql-general@postgre
I have a database which was created as LATIN1 (the machine has the
wrong locales installed when I set up PG). It is running 8.3.
So I found various places which said the way to do this was to do
a pg_dumpall -f dump_file, get rid of the entire database, init_db -E UTF-8,
and then psql -f dumpfi
12 matches
Mail list logo
