fft question about data length and spline data

I want use fft analizy my water level data.

because my data have lost ,so i use spline fill up my data.

but when i use data length exceed 8760 hr (365) day my spline line result not would be normal.

the attached file are the spline results of the test data length of one year and two years.

Answer 1

[Star Strider]

If you have R2016b or later, the fillmissing function will likely be a better choice.

[peter huang]

if my data have very large loss in a specific area

which way should i use fillmissing to missing values

attcached file is my raw-data and fillmissing result (spline)

[Star Strider]

I would not use the 'spline' nethod with fillmissing where there is a large gap. I would instead use 'linear' or some other choice that provides a reasonable approximation of the missing data. Another option is the Signal Processing Toolbox resample function.