How many cells you would want to assign values to?  If there are only a few, then maybe the conditional statement in ArcMap will do the trick.

In Spatial Analyst, open the Raster Calculator.   To replace the value in raster 1 with the values in raster 2, enter:

Con([Ras1] == value, [Ras2], [Ras1])

Click the link below for more.

http://support.esri.com/en/knowledgebase/techarticles/detail/34932

Hi Eric!

I tried with that function, however I need  fill the cells with a specific value, in this case  value of temperature from nearby cells. Basically I need to do something similar to expand function (spatial analysis tools/) in a FLOATING RASTER.

Cheers!!

Hey Ricardo,

dependent on how large your gaps are, and what kind of value you want to assign from the neighborhood (e.g. MEAN or MAXIMUM) you can use focal statistics. 

Running focal statistics only for the cell with NoData is easily possible using python. Therefore it is not realy necessary to write a scripts. In ArcMap you have the opportunity to use python commands.

Python command line can be activated by a button in the menu bar next to ArcToolbox. In the python command line write:

>>> from arcpy import*

>>> fill_gaps = sa.Con(sa.IsNull(in_raster_path), sa.FocalStatistics(in_raster_path, sa.NbrRectangle(w, h),'MEAN'), in_raster_path)

In this command : in_raster_path is the path of your raster dataset e.g. r'D:\WORKSPACE\test.tif'

fill_gaps is the output

w, h are  number of pixels for the width and hight of the rectangular neiborhood of the NoData value (this is what you have to Play with if you have larger regions with connected NoData pixels).

'MEAN' is the type of value that will be assigned to a NoData Pixel (it can be changed to 'MAJORITY', 'MAXIMUM', 'MEDIAN' etc. (the entire list can be seen exploring the focal statistics tool (Spatial Analyst Tools-> Neighborhood-> Focal statistics).

After you press enter and when the anylsis is ready the Analysis result will appear in your ArcMap document.

Best regards,

Jewgenij

Hi Ricardo,

Jewgenij gave you a good answer. Just if you don't want to use python (you should but who knows), if you use focal statistics, then you will have to mosaic the output raster and the original raster, using this one as target (and selecting  the "first" method in "mosaic_type" option). Be aware that the mosaic tool will modify the original raster, so you may want to use mosaic to new raster instead (or make a copy of the original before the mosaic).

Remember that focal statistics will work fine if you have isolated pixels with no data. Otherwise, it would be better to try something different like a spatial interpolation.

Cheers,

Jose

hi guys!!! thank very much for your advice!!