Thread: COPY Command UtfToLocal: could not convert UTF-8

COPY Command UtfToLocal: could not convert UTF-8

From
"kerdem"
Date:

I installed PG 7.3.4 and created DB with encoding “LATIN10”  . I tried to import Flat-File  with Copy Command

 

 

String  delimeter ;";

String  FILEPATH ="/home/kerdem/upstream/LIQ.txt";

String pgcopy ="copy public.liq_import from '"+FILEPATH+"' with delimiter '"+delimeter+"'  NULL AS '' ";

         pstmt = con.prepareStatement(pgcopy);

                        pstmt.executeUpdate(); 

                        pstmt.close();

 

I am getting  

 

WARNING: copy: line 2, UtfToLocal: could not convert UTF-8 (0xe47469). Ignored
WARNING: copy: line 7, UtfToLocal: could not convert UTF-8 (0xf66765). Ignored
WARNING: copy: line 7, UtfToLocal: could not convert UTF-8 (0xe4676e). Ignored
WARNING: copy: line 8, UtfToLocal: could not convert UTF-8 (0xe4747a). Ignored
WARNING: copy: line 12, UtfToLocal: could not convert UTF-8 (0xc472). Ignored
WARNING: copy: line 13, UtfToLocal: could not convert UTF-8 (0xe4727a).
Ignored
.......
....
..

I cannot import special letter like (ä,ß,ü,ö).

 

 

 

But  wenn I use same command  from psql console in DB  I can import all of the special letter,

 

  copy public.liq_import from '/home/kerdem/upstream/LIQ.txt' delimiter ';'  NULL as '' ;

 

I would appreciate any help.

 

 

Re: COPY Command UtfToLocal: could not convert UTF-8

From
Kris Jurka
Date:

On Wed, 10 Mar 2004, kerdem wrote:

> I installed PG 7.3.4 and created DB with encoding �LATIN10�  . I tried
> to import Flat-File  with Copy Command
>
> String  delimeter ;";
> String  FILEPATH ="/home/kerdem/upstream/LIQ.txt";
> String pgcopy ="copy public.liq_import from '"+FILEPATH+"' with
> delimiter '"+delimeter+"'  NULL AS '' ";
>
>          pstmt = con.prepareStatement(pgcopy);
>                         pstmt.executeUpdate();
>                         pstmt.close();
>
> WARNING: copy: line 2, UtfToLocal: could not convert UTF-8 (0xe47469).
> Ignored
>
> But  wenn I use same command  from psql console in DB  I can import all
> of the special letter,
>
>   copy public.liq_import from '/home/kerdem/upstream/LIQ.txt' delimiter
> ';'  NULL as '' ;
>

I imagine this problem is a result of the JDBC driver setting the
client_encoding variable to unicode (which it needs to function properly),
this is making the COPY command think the file has a unicode encoding.
When connecting with psql your client_encoding defaults to the database
encoding which I imagine is what the file is encoded as.

As to how to solve this problem?  It is possible to SET the
client_encoding temporarily and then reset it after the COPY is done, but
this is something I hesitate to recommend because it's definitely not
something the JDBC driver wants you doing.  I'm not sure what else to
could be done, extend the COPY command to take an encoding argument?

Kris Jurka