|
Hi,
If I do...
int nColName = odbcDataReader.GetOrdinal(@"name");
int nColNumber = odbcDataReader.GetOrdinal(@"iq");
while (odbcDataReader.Read())
{
string strName = odbcDataReader.GetString(nColNumber);
int nNumber = odbcDataReader.GetInt32(nColNumber);
}
GetString() succeeds.
GetInt32() throws an exception... [System.InvalidCastException] = {"Specified cast is not valid."}
However if I do...
int nColName = odbcDataReader.GetOrdinal(@"name");
int nColNumber = odbcDataReader.GetOrdinal(@"iq");
while (odbcDataReader.Read())
{
odbcDataReader.IsDBNull(nColNumber);
int nNumber = odbcDataReader.GetInt32(nColNumber);
string strNumber = odbcDataReader.GetString(nColNumber);
}
Now GetInt32() succeeds.
And GetString() throws an exception... [System.InvalidCastException] = {"Unable to cast object of type 'System.Int32' to type 'System.String'."}
FYI: The column in question 'IQ' is of type 'int'.
It seems that calling IsDBNull() first, changes the behaviour of the Get methods.
Can anyone explain what is going on?
Thanks - John.
|
|
|
|
|
My only guess at this point is in the first test you are calling GetString first, and in the second test you are calling GetInt32 first. It's possible the the reader is casting your value to what you asked for first?
|
|
|
|
|
Hi kevinnicol,
The only reason for putting them in a different order is because I put the call that throws the exception last.
Changing the order just means that you never get to second call because the first throws an exception.
@PIEBALDconsult: Yes, I should check the return but it makes no differance to my question.
|
|
|
|
|
Shouldn't it be like:
while (odbcDataReader.Read())
{
string strName = odbcDataReader.GetString(nColNumbernColName);
int nNumber = odbcDataReader.GetInt32(nColNumber);
}
__John_ wrote: @PIEBALDconsult: Yes, I should check the return but it makes no differance to my
question.
I beg to differ. If it's true e.g. the value is null it will throw an exception on the int part.
Ints can't be null. Either use a nullable int? or check for null first.
All the best,
Dan
|
|
|
|
|
Non of the returns are null so checking makes no fifferance.
|
|
|
|
|
Then yes. But you failed to specify that.
Honestly I'm out of ideas. Maybe a stupid question like "what is the DB type?" is all I got left.
Maybe it should be long or short or... on the C# side.
All the best,
Dan
|
|
|
|
|
Hi MDL=>Moshu, thanks for the reply.
The cast that fails is internal to the Get methods. So for example...
var Val = odbcDataReader.GetInt32(nColNumber);
Still thows the same exception.
I am using the data provider for MS SQL Sever that came with vs2010.
SQL Server version is 2008.
Thanks - John.
|
|
|
|
|
In your statement var Val is the same as int Val cause GetInt32() returns a int on success.
var Val = odbcDataReader[nColNumber]; is what you should use.
Val will be an object of type Object and you can check for null and do the casting later.
This is obviously for debugging purposes. Cause the cast and checking are not necessarily once you figure
out the problem.
All the best,
Dan
|
|
|
|
|
Hi Dan,
Using the indexer, casting to an int always works and casting to a string always fails (invalid cast).
So the behavour is predictable and my problem 'kind of' goes away. So for now I will use the indexer and not use the Get???? methods.
@PIEBALDconsult: You mentioned indexers earlier but I was not sure what you were getting at. Turns out you were on to something. Thanks.
|
|
|
|
|
Again you used nColNumber twice. You should have used nColNumber and nColName.
That was your problem. And yeah I don't use GetXXX either. But that wasn't what caused the exception.
You tried to get a string from an int DB value or a int form a varchar DB value courtesy
of using twice nColNumber.
All the best,
Dan
|
|
|
|
|
Well, personally, I never use the Get methods anyway; I just use the indexers.
__John_ wrote: odbcDataReader.IsDBNull(nColNumber);
Shouldn't you be testing the result?
|
|
|
|
|
What is the query that gets executed on the DB while doing a cmd.ExecuteReader()?
Maybe there is something missing/wrong there.
All the best,
Dan
|
|
|
|
|
Hi Moshu,
The command that I execute is as follows..
odbcSelectCommand.CommandText = @"SELECT name, iq FROM Table_People WHERE iq < 3 ORDER BY iq";
I get the 'name' and 'iq' but I am only concurned with 'iq' for now.
|
|
|
|
|
OK. Heres what I would do. Run that query in SQLManagementStudio and check the results.
Are there any? Or some with a field null or...
All the best,
Dan
|
|
|
|
|
Hi MDL=Moshu, thanks for looking at this.
The code now looks like this...
int nColName = odbcDataReader.GetOrdinal(@"name");
int nColNumber = odbcDataReader.GetOrdinal(@"iq");
while (odbcDataReader.Read())
{
if(odbcDataReader.IsDBNull(nColNumber))
{
continue;
}
int nNumber = odbcDataReader.GetInt32(nColNumber);
string strNumber = odbcDataReader.GetString(nColNumber);
}
No nulls are returned and the behaviour is the same ie. with the call to IsDBNull() GetString() fails, and without it, GetInt32() fails.
|
|
|
|
|
This is my final reply. First check my post below.
Second you're using nColNumber for both the name and the iq.
That's not right and I pointed it out on my first post to this thread.
If for some reason you want a string representation of the number/iq then
just use strNumber = nNumber.ToString();
Hope it helps. I'm out of CP for today. Going to program some beers now.
All the best,
Dan
|
|
|
|
|
Also since you are not using a Select * statement, I would drop the lines of code that get the ordinals.
You know them name is 0 and iq is 1 so in a nutshell:
while(dr.Read()){
string name = dr[0].ToString();
int iq = (int)dr[1];
}
All the best,
Dan
|
|
|
|
|
Thanks dan,
But I think my way is more robust and future proof.
ie. it will still work even with 'SELECT *'.
- John
|
|
|
|
|
True. Then use dr["name"] and dr["iq"] and it will be future proof.
All the best,
Dan
|
|
|
|
|
Hi MDL=>Moshu,
Using an indexer as you sagest (with column name) is fine if you only have a small number of fields, but if the number of fields is higher then I think you will take a performance hit.
|
|
|
|
|
__John_ wrote: my way is more robust
That doesn't sound very convincing as your code does not produce the expected results to begin with.
Have you already figured nColNumber is appearing too many times in your code? (courtesy MDL)
|
|
|
|
|
MDL=>Moshu wrote: string name = dr[0].ToString();
It's already a string, just cast it -- string name = (string) dr[0]; , no need to call a method.
|
|
|
|
|
True.
Is there a performance reason too, or just style?
Cause if there is one, I might need to change some lines of code. I've always happily used ToString() on
datareader where the element is a text/string.
All the best,
Dan
|
|
|
|
|
There must be at least a small overhead in calling the method.
|
|
|
|
|
Your code changed. When you test things you should change 1 item at a time, however, you have also changed order and type. In your last example, you are calling GetString and GetInt32 on nColNumber which refers to an integer, your code should read:
int nameColumn = odbcDataReader.GetOrdinal(@"name");
int idColumn = odbcDataReader.GetOrdinal(@"iq");
while (odbcDataReader.Read()){
odbcDataReader.IsDBNull(idColumn);
int iqValue= odbcDataReader.GetInt32(idColumn);
string nameValue = odbcDataReader.GetString(nameColumn );
}
I just typed it in the post window but you can see how different variable names make it easier to see.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
