standardizeMissing not converting selected values to NaN

I'm having trouble getting standardizeMissing to perform its function. I have a 3x4209 array of int32 values called bl_height. Some of the values are -999. I'm trying to change the -999 values to NaN using standardizeMissing, but the -999 values are retained. Can someone tell me what I'm doing wrong? I've tried pretty much every iteration of standardizeMissing possible. Here's my code below:

bl_height=ncread(filename,'bl_height');

bl_height_edited = standardizeMissing(bl_height,-999);

Answer 1

[Taylor]

It's because your values are integers not doubles. There is no room for a NaN value when using integers. The documentation for the input data to standardizeMissing specifies that the input should be one of the following data types: double | single | char | string | cell | table | timetable | categorical | datetime | duration

https://www.mathworks.com/help/matlab/ref/standardizemissing.html#bty9pdr-1-A

[Christopher Traft]

Thanks, it worked after I converted the values to doubles. Changed my first line to:

bl_height = double(ncread(filename,'bl_height'));

[Taylor]

Great! Just be aware that if you try to convert back to integers the NaN values will become 0s.