Change raster `nodata` to nan on python

Question

I've been trying to use rasterio.merge.merge() but it seems I have a nodata issue forbidding it (Rasterio.merge.merge() does not work): there is no value defined as nodata for my rasters (they all have the same metadata). I would like to change it to nan in format float32. How could I do this, on Python ?

The metadata of the raster is:

print('Raster 1: \n', raster1.profile)
Raster 1: 
 {'driver': 'GTiff', 'dtype': 'float32', 'nodata': None, 'width': 3532, 'height': 3537, 'count': 16, 'crs': CRS.from_epsg(32607), 'transform': Affine(30.0, 0.0, 484200.0,
       0.0, -30.0, 6719670.0), 'blockxsize': 256, 'blockysize': 256, 'tiled': True, 'compress': 'lzw', 'interleave': 'pixel'}

And the format of one nan cell is:

Answer 1

Open the dataset in append mode "r+" and then set dataset.nodata (source):

with rio.open(src_filename, "r+") as src:
       src.nodata = np.nan

For example:

import numpy as np
import rasterio as rio

src_filename = '/path/to/some.tif'

with rio.open(src_filename) as src:
       print(src.read(masked=True))
       print(src.nodata)

with rio.open(src_filename, "r+") as src:
       src.nodata = np.nan
       print(src.read(masked=True))
       print(src.nodata)

Output:

[[[nan nan  3. ...  4. nan nan]
  [nan  3. nan ... nan  5. nan]
  [ 3. nan  4. ... nan nan  4.]
  ...
  [ 6. nan nan ...  5.  3.  3.]
  [ 4. nan  6. ...  6.  5.  5.]
  [nan  4. nan ...  6.  4.  3.]]]
None
[[[-- -- 3.0 ... 4.0 -- --]
  [-- 3.0 -- ... -- 5.0 --]
  [3.0 -- 4.0 ... -- -- 4.0]
  ...
  [6.0 -- -- ... 5.0 3.0 3.0]
  [4.0 -- 6.0 ... 6.0 5.0 5.0]
  [-- 4.0 -- ... 6.0 4.0 3.0]]]
nan