Hi all, We support binning in transform encode and NaN if we try to encode values not in any Bin.
e.g. if the bins are : 0-1,1-2,2-3,3-4,4-5,5-6,6-7,7-8,8-9,9-10 If we try to encode 0.1 we return 1 for bin 1, or 8.8 we return 9. The issue I would like to address comes with values close to the boundaries where potential rounding errors in double values make us return Nan on inputs like 10.0000000001. I suggest that we open the end boundaries to encode any value that is above or below the end buckets like: -inf-1,1-2,2-3,3-4,4-5,5-6,6-7,7-8,8-9,9-inf This meanse that -13 is in bucket 1 and 1324 is in bucket 10, while we keep NaN mapping to NaN. Best regards Sebastian
