Oups sorry. Always read twice the documentation... Indeed I read in the FNR transform paragraph :"In the case of points being equally spaced around the missing point, the mean value is used".
I was confused because in the example I was working on, interpolation has no sense. Imagine a type of photoplancton, an interpolation brings a specie that was not initially present.
I was really expecting a nearest filled even in the case of equally spaced.Anyway the problem is there when an axis irregular is set (second part of my mail).
Patrick
Hi, Why the filled nearest transform does not find the nearest value ? Try: yes? let A={4,-1E+34,6} yes? list A,A[i=@FLN] X: 0.5 to 3.5 Column 1: A is {4,-1E+34,6} Column 2: A[X=@FLN:1] is {4,-1E+34,6} (linear-filled by 1 pts on X) A A 1 / 1: 4.000 4.000 2 / 2: .... 5.000 3 / 3: 6.000 6.000 It should be 4 or 6 not 5 which is a interpolation value ? Same thing if the variable A is set along an axis yes? def axis/x myx1={1,2,5} yes? let B=A[gx=myx1@asn] yes? list B, B[x=@FNR] X: 0.5 to 6.5 Column 1: B is A[GX=MYX1@ASN] Column 2: B[X=@FNR:1] is A[GX=MYX1@ASN] (nearest-filled by 1 pts on X) B B 1 / 1: 4.000 4.000 2 / 2: .... 5.000 5 / 3: 6.000 6.000I think there is a problem and I suspect that FNR and FLN act the same way.Can this be confirmed ? Tested with ferret v6.61 Patrick
-- LSCE/IPSL, Laboratoire CEA-CNRS-UVSQ Data Analysis and Visualization Engineer IPSL Global Climate Modelling Group mailto:Patrick.Brockmann@xxxxxx 01.69.08.32.24 --> LSCE (Orme des merisiers 712, Bureau 11) http://www.ipsl.jussieu.fr/~brocksce/ --