Thread: 08.02.04.00 Updating column with Unicode

Hello everybody,

I have detected some strange behavior updating columns with Unicodes strings.

Environment:
PostgreSQL 8.2.5
Database encoding: UTF8
ODBC: PostgreSQL Unicode Version 08.02.04.00

I am updating via ODBC a column (character varying nnn) of a record to an unicoded string.

Updating by using the command as one parameterblock
'UPDATE adresse SET company = ''Seifert und Partner KG € 文件'' WHERE id = 1'
works right. (the whole parameterblock is encoded to UTF8)

Updating the record by using the command 'UPDATE adresse SET company=? WHERE record.id=?'
and providing the arguments for ? UTF8-encoded in an extra buffer destroys the columns content.

If i update the column with an UTF8 encoded character in the default range (the EURO Sign €), it works.

Can anybody give me some hints ?

mit freundlichen Grüßen / best regards,
Josef Springer
(Geschäftsleitung/Management)

	Address	Josef.Springer@joops.com	Orlando-di-Lasso Str. 2	D-85640 Putzbrunn
	Office	+49 (0)89 600 6920
	Fax	+49 (0)89 600 69220
	Web	http://www.joops.com

(HRB München 86239) -- the software company --

(2011/04/03 20:50), Josef Springer wrote:
> Hello everybody,
>
> I have detected some strange behavior updating columns with Unicodes
> strings.
>
> Environment:
> PostgreSQL 8.2.5
> Database encoding: UTF8
> ODBC: PostgreSQL Unicode Version 08.02.04.00
>
> I am updating via ODBC a column (character varying nnn) of a record to
> an unicoded string.
>
> Updating by using the command as one parameterblock
> 'UPDATE adresse SET company = ''Seifert und Partner KG € 文件'' WHERE id
> = 1'
> works right. (the whole parameterblock is encoded to UTF8)
>
> Updating the record by using the command 'UPDATE adresse SET company=?
> WHERE record.id=?'
> and providing the arguments for ? UTF8-encoded in an extra buffer
> destroys the columns content.
>
> If i update the column with an UTF8 encoded character in the default
> range (the EURO Sign €), it works.
>
> Can anybody give me some hints ?

How do you bind the parameters?

regards,
Hiroshi Inoue

Hi Hiroshi,

the parameters are bound as an UTF8-encoded bytearray. Because the interface is controlled by a Smalltalk application, i think the codings does not help you:
bytes := v asByteArrayEncoding: #'utf-8'.
p copyAt: 0 from: bytes size: (size := bytes size) startingAt: 1.
p at: size put: 0

Josef Springer

Hiroshi Inoue wrote:

(2011/04/03 20:50), Josef Springer wrote:

Hello everybody,

I have detected some strange behavior updating columns with Unicodes
strings.

Environment:
PostgreSQL 8.2.5
Database encoding: UTF8
ODBC: PostgreSQL Unicode Version 08.02.04.00

I am updating via ODBC a column (character varying nnn) of a record to
an unicoded string.

Updating by using the command as one parameterblock
'UPDATE adresse SET company = ''Seifert und Partner KG € 文件'' WHERE id
= 1'
works right. (the whole parameterblock is encoded to UTF8)

Updating the record by using the command 'UPDATE adresse SET company=?
WHERE record.id=?'
and providing the arguments for ? UTF8-encoded in an extra buffer
destroys the columns content.

If i update the column with an UTF8 encoded character in the default
range (the EURO Sign €), it works.

Can anybody give me some hints ?

How do you bind the parameters?

regards,
Hiroshi Inoue

Hi Josef,

(2011/04/04 22:47), Josef Springer wrote:
> Hi Hiroshi,
>
> the parameters are bound as an UTF8-encoded bytearray. Because the
> interface is controlled by a Smalltalk application, i think the codings
> does not help you:
> /bytes := v asByteArrayEncoding: #'utf-8'.
> p copyAt: 0 from: bytes size: (size := bytes size) startingAt: 1.
> p at: size put: 0
> /

Unfortunately I know nothing about Smalltalk.
Anyway *Unicode* means UTF-16 as for the drivers.
Could you try to give utf-16 strings?

regards,
Hiroshi Inoue

> Josef Springer
>
> Hiroshi Inoue wrote:
>> (2011/04/03 20:50), Josef Springer wrote:
>>> Hello everybody,
>>>
>>> I have detected some strange behavior updating columns with Unicodes
>>> strings.
>>>
>>> Environment:
>>> PostgreSQL 8.2.5
>>> Database encoding: UTF8
>>> ODBC: PostgreSQL Unicode Version 08.02.04.00
>>>
>>> I am updating via ODBC a column (character varying nnn) of a record to
>>> an unicoded string.
>>>
>>> Updating by using the command as one parameterblock
>>> 'UPDATE adresse SET company = ''Seifert und Partner KG € 文件'' WHERE id
>>> = 1'
>>> works right. (the whole parameterblock is encoded to UTF8)
>>>
>>> Updating the record by using the command 'UPDATE adresse SET company=?
>>> WHERE record.id=?'
>>> and providing the arguments for ? UTF8-encoded in an extra buffer
>>> destroys the columns content.
>>>
>>> If i update the column with an UTF8 encoded character in the default
>>> range (the EURO Sign €), it works.
>>>
>>> Can anybody give me some hints ?
>>
>> How do you bind the parameters?
>>
>> regards,
>> Hiroshi Inoue

Hello Hiroshi,

the encoding of the databsase is set to UTF-8. So i think, i must use
UTF-8. I have tried allready giving UTF-16 strings, but with the same
results.

Thanks for your help,
Josef

Hiroshi Inoue wrote:
> Hi Josef,
>
> (2011/04/04 22:47), Josef Springer wrote:
>> Hi Hiroshi,
>>
>> the parameters are bound as an UTF8-encoded bytearray. Because the
>> interface is controlled by a Smalltalk application, i think the codings
>> does not help you:
>> /bytes := v asByteArrayEncoding: #'utf-8'.
>> p copyAt: 0 from: bytes size: (size := bytes size) startingAt: 1.
>> p at: size put: 0
>> /
>
> Unfortunately I know nothing about Smalltalk.
> Anyway *Unicode* means UTF-16 as for the drivers.
> Could you try to give utf-16 strings?
>
> regards,
> Hiroshi Inoue
>
>> Josef Springer
>>
>> Hiroshi Inoue wrote:
>>> (2011/04/03 20:50), Josef Springer wrote:
>>>> Hello everybody,
>>>>
>>>> I have detected some strange behavior updating columns with Unicodes
>>>> strings.
>>>>
>>>> Environment:
>>>> PostgreSQL 8.2.5
>>>> Database encoding: UTF8
>>>> ODBC: PostgreSQL Unicode Version 08.02.04.00
>>>>
>>>> I am updating via ODBC a column (character varying nnn) of a record to
>>>> an unicoded string.
>>>>
>>>> Updating by using the command as one parameterblock
>>>> 'UPDATE adresse SET company = ''Seifert und Partner KG € 文件''
>>>> WHERE id
>>>> = 1'
>>>> works right. (the whole parameterblock is encoded to UTF8)
>>>>
>>>> Updating the record by using the command 'UPDATE adresse SET company=?
>>>> WHERE record.id=?'
>>>> and providing the arguments for ? UTF8-encoded in an extra buffer
>>>> destroys the columns content.
>>>>
>>>> If i update the column with an UTF8 encoded character in the default
>>>> range (the EURO Sign €), it works.
>>>>
>>>> Can anybody give me some hints ?
>>>
>>> How do you bind the parameters?
>>>
>>> regards,
>>> Hiroshi Inoue
>
>

Hi Josef,

Could you send me directly the Mylog output?

regards,
Hiroshi Inoue

(2011/04/05 17:27), Josef Springer wrote:
> Hello Hiroshi,
>
> the encoding of the databsase is set to UTF-8. So i think, i must use
> UTF-8. I have tried allready giving UTF-16 strings, but with the same
> results.
>
> Thanks for your help,
> Josef