|
| Hi everyone,
In debugging a problem we have when the SQL Server goes offline, I found
that the behaviour of SQLFetch is not what I expected.
I have code which calls SQLExecDirect and then followed by SQLFetch. Now to
simulate the problem I run my application in debug mode and run past the
SQLExecDirect call but breaking at the SQLFetch call.
For the query that I'm running, it should return about 800 rows. When I
reach the breakpoint (at SQLFetch), I stop the SQL Server service -
effectively simulating a disconnection. When I trace through my SQLFetch
loop, I find that it returns about 16 rows before returning SQL_NO_DATA. I
would have imagined the first call would have returned SQL_ERROR. Even if SQL
Server did buffer some of its return data, I would have thought after the
16th row it would return SQL_ERROR rather than the misleading SQL_NO_DATA. Am
I right to expect this or is there a better way to detect disconnections?
Any help or insights much appreciated!
|
|