Is there a difference in performance when it comes to working with Missing vs Null values in FME?
I'm working with some database tables which have a lot of null values. The output is something like JSON so In the end I want the null values to be missing in the output rather than a null or an empty string.
but when it comes for manipulating the data is there a difference between missing and null values?
Not a definative answer, but my thought process would be that missing is faster as at the feature schema level, it doesn't exist, fewer fields, less data, faster processing....
But I might be wrong!
Not a definative answer, but my thought process would be that missing is faster as at the feature schema level, it doesn't exist, fewer fields, less data, faster processing....
But I might be wrong!
Yeah thanks Hamish - My thinking was that with Bulk Mode it's better to have all the features share the same table structure. I'm not sure if having a lot of missing attributes messes that up. Any ideas @mark2atsafe ?
I did some minor testing but results were inconclusive
Yeah thanks Hamish - My thinking was that with Bulk Mode it's better to have all the features share the same table structure. I'm not sure if having a lot of missing attributes messes that up. Any ideas @mark2atsafe ?
I did some minor testing but results were inconclusive
I asked around, and we don't really think there would be a huge difference. You'd need to have hundreds of null/empty attributes per feature to even make it worth testing. But if I hear different from another source, I'll let you know.
Yeah thanks Hamish - My thinking was that with Bulk Mode it's better to have all the features share the same table structure. I'm not sure if having a lot of missing attributes messes that up. Any ideas @mark2atsafe ?
I did some minor testing but results were inconclusive
Awesome thanks mark. definitely not hundreds - more in the dozens. Seems to be working well either way so I'm happy.