>Hi, as far as I've understood client_encoding tells postgresql
>how data "will arrive / must be send back" to the client application.
>Postgresql will do the conversion for you if the database encoding is different.
>
>I've a unicode database and the line
>client_encoding=latin1 in the postgresql configuration file
>(this is the same as calling set client_encoding='latin1' on any connection right?).
>
>i insert one row containing the Hex 92 characther using the npgsql net driver :
>I sniff the wire and i see that that only the hex 92 bvbyte is sent.
>
>i insert one row containing the Hex 92 characther using the npgsql net driver
>setting Encoding=UNICODE in the cnstring:
>I sniff the wire and i see that that 3 bytes are sent : the coorect equivalent in unicode (i see that a set client_ecoeding=UNICODE is sent).
>
>Now i want to read the data using :
>1) if i read the first row keeping client_encoding latin1 everything is OK (i get hex 92 back)
>2) if i read the 2nd row setting client_encoding unicode everything is OK (i get the 3 unicode bytes)
>3) if i read the first row setting client_encoding unicode IT DOES NOT WORK I get a 2 byte sequence c2 92 ?? so i see garbage.
>
>WHY this does not work (can't i write in one encoding and read in another?)
>Looks like postgresql did not correctly converted from latin1 to unicode and has gargabage inside
>(but why do i read correctly the first line when i have client_encoding latin1) ?
>
>PLEASE help
>
>best regards
>Enrico Sabbadin
>
>
>
>
>
>
>
>
>
>
