Unicode parametric query

[djdjdjole]

I don't know whether this question is for Lazarus or UniDAC forum. I have some Serbian Latin strings in my SQL Server. My query is simple one and is quoted in UniQuery SQL property at design time as: select * from mytable where name='Šeki'. I correctly receive the only record which satisfied condition. However when I change the query to: select * from mytable where name=:par0 and in code: query.parambyname(par0).asstring:= 'Šeki', I receive no record. Field "name" is "nvarchar" in database, collation Croatian, client OS- Windows10 (64- bit), database machine is Windows Server 2012 (64- bit) and Lazarus version is 1.6.

[zeljko]

You written everything except fpc version and database type, anyway, I'm using postgresql + ZEOS + Lazarus trunk (QtLCL) + fpc 3.0.0 (2.6.4 also worked fine) and never had such problems (linux, mac, windows). eg
RQRY = zeos readonly query
RQRY.SQL.Add('SELECT * FROM names WHERE name=:P1');
RQRY.Params[0].AsString := 'Željko';
RQRY.Active := True;
.... works fine on any mentioned OS ...returns record

[djdjdjole]

FPC is 3.0 and database is SQL Server 2012 Express. No matter I will try my luck with Zeos components, but still sounds strange to me that UniDAC (I understood that you pointed the finger to it) behaves differently in two mentioned situations. 

[paweld]

Code: Pascal  [Select][+][-]

1. Query.SQL.Text:='select * from mytable where name=:par0';
2. //if "name" field type is varchar
3. Query.Params[0].AsString:=UTF8ToCP1250('Šeki'); //add LConvEncoding unit to uses section
4. //or if "name" field type is nvarchar
5. Query.Params[0].AsString:=UTF8ToUTF16('Šeki'); //add LazUTF8 unit to uses section

[djdjdjole]

I've already tried that with no success. It sounds reasonable, because SQL Server's encoding format is UTF-16 (as much as I know) and Lazarus's one is UTF-8, but the problem is obviously somewhere else.

[djdjdjole]

My excuse to paweld - I didn't noticed line number 3, which I didn't try . It should be first thing tomorrow.

[djdjdjole]

The field is of "nvarchar" type in database and attempt with Utf8ToUtf16 doesn't work- returns no record. When building SQL query in source :
SQL.Add('select * from myTable where name=');
SQL.Add(QuotedStr('Šeki'));
it does work- returns one record.

[zeljko]

Have you tried to add {$codepage utf8} into that unit ? Maybe 'Šeki' is interpreted as cp1250 ...

[djdjdjole]

Logical assumption, still it gives no solution. So I decided to build query in code- without parameters.

[paweld]

try to use macros:

Code: Pascal  [Select][+][-]

1. Query.SQL.Text:='select * from mytable where name=&mac0';
2. Query.Macros[0].AsString:='Šeki';

[djdjdjole]

Simply no way- tried and returned no record. You guys are so kind and persistent - I have no words to express my gratitude. At the moment I found the way to override the problem (I already wrote about), but it still remains mystery to me.