Dynamic type casting at run time

[Earl Teigrob]

I am using reflection to iterate the fields of a class and need to cast
these (sqldata) types to there .net primitive counterparts. The code I am
currently using for this is certainly a brute force method but it does work.
Is there a more elegant way of doing this? (See code below)

Thanks

Earl

string p;
foreach (System.Reflection.FieldInfo myFieldInfo in
DataRecord.GetType().GetFields())
{
object x = myFieldInfo.GetValue(DataRecord);
if (!SqlTypeIsNull(x))
{
p=x.GetType().ToString();
switch (p)
{
case "System.Data.SqlTypes.SqlBoolean":
r[myFieldInfo.Name]=(Boolean)(System.Data.SqlTypes.SqlBoolean)x;
break;
case "System.Data.SqlTypes.SqlByte":
r[myFieldInfo.Name]=(Byte)(System.Data.SqlTypes.SqlByte)x;
break;
case "System.Data.SqlTypes.SqlDateTime":
r[myFieldInfo.Name]=(DateTime)(System.Data.SqlTypes.SqlDateTime)x;
break;
case "System.Data.SqlTypes.SqlDecimal":
r[myFieldInfo.Name]=(Decimal)(System.Data.SqlTypes.SqlDecimal)x;
break;
case "System.Data.SqlTypes.SqlDouble":
r[myFieldInfo.Name]=(Double)(System.Data.SqlTypes.SqlDouble)x;
break;
case "System.Data.SqlTypes.SqlInt16":
r[myFieldInfo.Name]=(Int16)(System.Data.SqlTypes.SqlInt16)x;
break;
case "System.Data.SqlTypes.SqlInt32":
r[myFieldInfo.Name]=(Int32)(System.Data.SqlTypes.SqlInt32)x;
break;
case "System.Data.SqlTypes.SqlInt64":
r[myFieldInfo.Name]=(Int64)(System.Data.SqlTypes.SqlInt64)x;
break;
case "System.Data.SqlTypes.SqlSingle":
r[myFieldInfo.Name]=(Single)(System.Data.SqlTypes.SqlSingle)x;
break;
case "System.Data.SqlTypes.SqlString":
r[myFieldInfo.Name]=(String)(System.Data.SqlTypes.SqlString)x;
break;
default:
r[myFieldInfo.Name]=(String)(System.Data.SqlTypes.SqlString)x;
break;
}
}
}

 

[Michael Culley]

Why do you need to cast them only to put them back into an object?

p=x.GetType().ToString();

Best not to use a string here, you've just thrown any sort of compiler checking out the window.

case "System.Data.SqlTypes.SqlString":
r[myFieldInfo.Name]=(String)(System.Data.SqlTypes.SqlString)x;
break;
default:
r[myFieldInfo.Name]=(String)(System.Data.SqlTypes.SqlString)x;
break;

Why duplicate this?

 

[Earl Teigrob]

Michael,

Thanks for your response...and well, thats why a make this post, to give me
a better way of doing this.. so can you give me an example of what you are
taking about?

Thanks

Earl

Why do you need to cast them only to put them back into an object?

p=x.GetType().ToString();

Best not to use a string here, you've just thrown any sort of compiler checking out the window.

case "System.Data.SqlTypes.SqlString":
r[myFieldInfo.Name]=(String)(System.Data.SqlTypes.SqlString)x;
break;
default:
r[myFieldInfo.Name]=(String)(System.Data.SqlTypes.SqlString)x;
break;

Why duplicate this?

 

[Michael Culley]

Michael,

Thanks for your response...and well, thats why a make this post, to give me
a better way of doing this.. so can you give me an example of what you are
taking about?

I don't think what you are doing will have any affect, eg

object X = true;
object Y = (bool) X;

will be the same as just

object X = true;
object Y = X;

You don't need the type cast because it is going back into an object, although I don't understand your double type cast, it seems to
cast it to an enum and then a string?:

=(String)(System.Data.SqlTypes.SqlString)

checking out the window.

You would be better using:

Type myType = X.GetType();
if (myType == typeof(sqlTypes.Int)) DoSomething;
etc

that way if you mistype sqlTypes.Int (as I probably have) then the compiler will tell you.

 

[Jon Skeet [C# MVP]]

You don't need the type cast because it is going back into an object,
although I don't understand your double type cast, it seems to
cast it to an enum and then a string?:

=(String)(System.Data.SqlTypes.SqlString)

That's because the object itself is a SqlString, which can then be
explicitly converted to a String. It looks odd, but it *does* actually
make a difference.

You would be better using:

Type myType = X.GetType();
if (myType == typeof(sqlTypes.Int)) DoSomething;
etc

that way if you mistype sqlTypes.Int (as I probably have) then the
compiler will tell you.

True - but you end up with a nasty big if/else block instead of a
switch block. Type.GetTypeCode provides a solution which has the best
of both worlds.

 

[Earl Teigrob]

Thanks, Jon

Yes, I must use the double cast for this to work.

I will use Type.GetTypeCode so that the case statement does not have to
evaluate a big long string.

I wish C# supported dynamic runtime casting so that I could cast based on a
calculation instead of a lookup, but the way I work around this issue is
livable...

Thanks for your help

Earl