07-05-2007 04:23 PM
07-06-2007 10:05 AM
@unknown00 wrote:
... is it not possible to delete the cluster double?
First of all, the floating point number in your cluster is a SGL and not a DBL. As mike already mentioned, you cannot surgically change the cluster type for certain array elements, so if you have a scalar SGL, it needs a value. There is no special symbol for "no value" or "missing value", for example but sometimes you can substitute NaN if you later remember what NaN means in the context of the code. This has been one of the early answere here.
As I said above, you could change your cluster type so it contains an array of SGL with lenght=1 containing your desired number and then make the array size zero for clusters where you don't want it. However, this seems a bit too convoluted and will overly complicate your data structures
It took me a long time to understand that with "double null" you actually wanted a "null" with datatype DBL. I thought you wanted a single byte (string or U8) containing the hex value value x00 (If you would come from a German speaking background, a null is a zero and we literally see a double zero) or maybe a 2 byte datatype with two null character bytes joined. 😄