Problem with regrid operation


#1

Hello. I’m using SciDB 18.1 on CentOS 7.4
My original array size is 1 X 12214 X 10082. Each array cell has 3 attributes (RGB data).
I need to reduce the picture resolution to be 640 X 480, so I use regrid operator with parameters 1, 19, 21.
I assume that 12214/19 is about 640 and 10082/480 is about 480.
But when I apply regrid operator by SciDBPy, the result array size is 6417 X 4793, which is 10 times higher in each dimension as expected.
The query looks like this:

result_array = db_obj.sort(db_obj.apply(db_obj.regrid(db_obj.between(db_obj.arrays[array_name], ‘null’, ‘null’, ‘null’, ‘null’, ‘null’, ‘null’), 1, x_regrid, y_regrid, ‘avg(red) as newRed, avg(green) as newGreen, avg(blue) as newBlue’), ‘latitude_i’, ‘int64(latitude)’, ‘longitude_i’, ‘int64(longitude)’), ‘latitude_i, longitude_i’)

Please help. Why regrid operation reduces sizes 10 times less than requested?
Thank you.


#2

Hi @igorch,
I think I can infer what your array schema is, but for clarity can you please post it explicitly? Also, I see the call to regrid in your python snippet, but I don’t see the parameters 1, 19, and 21 there. Please double-check the snippet for me. Please see the regrid operator reference page here: https://paradigm4.atlassian.net/wiki/x/CIgpK
Thanks!
Dave