CODEWITHSUNDEEP
pythonnumpyscipy
Weiss
Weiss

Reputation: 303

numpy.core._exceptions.UFuncTypeError: Cannot cast ufunc 'subtract' output from dtype('float64') to dtype('int32') with casting rule 'same_kind'

I want to create a grid data, which I plot on a map, the lat maximum of the grid should be at +90 and the minimum at -90! however, I keep getting the following error:

numpy.core._exceptions.UFuncTypeError: Cannot cast ufunc 'subtract' output from dtype('float64') to dtype('int32') with casting rule 'same_kind'

I have already tried:

from traceback import print_tb
import numpy as np
import pandas as pd
def get_data():
       "lats": np.array([6.47, 4.8, 1.94 ])
       "lons": np.array([46.37, 43.9, 47.83)
       "values": np.array([-29.5, -27.6, -32.5  ])
    }
def extend_data(data):
    return {
        "lons": np.concatenate([np.array([lon-360 for lon in data["lons"]]), data["lons"], np.array([lon+360 for lon in data["lons"]])]),
        "lats":  np.concatenate([data["lats"], data["lats"], data["lats"]]),
        "values":  np.concatenate([data["values"], data["values"], data["values"]]),
    }

def generate_grid(data, basemap, delta=1):
    latmax = 90
    latmin = -90
    latmax = latmax.astype('float32')
    latmin = latmin.astype('float32')
    grid = {
        'lon': np.arange(-180, 180, delta),
        'lat': np.arange(latmin, latmax, delta) 
    }
    grid["x"], grid["y"] = np.meshgrid(grid["lon"], grid["lat"])
    grid["x"], grid["y"] = basemap(grid["x"], grid["y"])
    return grid

base_data = get_data()
figure, axes, basemap = prepare_map_plot()
grid = generate_grid(base_data, basemap, 1)
extended_data = extend_data(base_data)

The error occours in line: extended_data = extend_data(base_data)

However when I write:

np.amin(data["lats"]), np.amax(data["lats"])

instead of:

    latmax = 90
latmin = -90
latmax = latmax.astype('float32')
latmin = latmin.astype('float32')
grid = {
    'lon': np.arange(-180, 180, delta),
    'lat': np.arange(latmin, latmax, delta)

The error dosn´t occour, however this is not what I wanted!

But this does not work, what could be the reason?

Upvotes: 1

Views: 3820

Answers (3)

hpaulj
hpaulj

Reputation: 231385

syntax errors in:

def get_data():
       "lats": np.array([6.47, 4.8, 1.94 ])
       "lons": np.array([46.37, 43.9, 47.83)
       "values": np.array([-29.5, -27.6, -32.5  ])
    }

Missing function: prepare_map_plot

After correcting get_data:

In [146]: base_data = get_data()

In [147]: base_data
Out[147]: 
{'lats': array([6.47, 4.8 , 1.94]),
 'lons': array([46.37, 43.9 , 47.83]),
 'values': array([-29.5, -27.6, -32.5])}

extended_data runs:

In [149]: extended_data = extend_data(base_data)

In [150]: extended_data
Out[150]: 
{'lons': array([-313.63, -316.1 , -312.17,   46.37,   43.9 ,   47.83,  406.37,
         403.9 ,  407.83]),
 'lats': array([6.47, 4.8 , 1.94, 6.47, 4.8 , 1.94, 6.47, 4.8 , 1.94]),
 'values': array([-29.5, -27.6, -32.5, -29.5, -27.6, -32.5, -29.5, -27.6, -32.5])}

Rewriting generate to skip the unavailable basemap

In [151]: def generate_grid(data, delta=1):
     ...:     latmax = 90
     ...:     latmin = -90
     ...:     latmax = latmax.astype('float32')
     ...:     latmin = latmin.astype('float32')
     ...:     grid = {
     ...:         'lon': np.arange(-180, 180, delta),
     ...:         'lat': np.arange(latmin, latmax, delta) 
     ...:     }
     ...:     grid["x"], grid["y"] = np.meshgrid(grid["lon"], grid["lat"])
     ...:     # grid["x"], grid["y"] = basemap(grid["x"], grid["y"])
     ...:     return grid
     ...:

Note that the basemap line overwrites the previous meshgrid variables.

In [152]: generate_grid(base_data)
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Input In [152], in <cell line: 1>()
----> 1 generate_grid(base_data)

Input In [151], in generate_grid(data, delta)
      2 latmax = 90
      3 latmin = -90
----> 4 latmax = latmax.astype('float32')
      5 latmin = latmin.astype('float32')
      6 grid = {
      7     'lon': np.arange(-180, 180, delta),
      8     'lat': np.arange(latmin, latmax, delta) 
      9 }

AttributeError: 'int' object has no attribute 'astype'

I get an error, but it's unrelated to the one you get. You did not post the code that actually produced your error, and you did not post the full error.

I'm going to vote to close this question because the it needs proper debugging details.

Upvotes: 1

Tae In Kim
Tae In Kim

Reputation: 224

Unsafe casting will do the operation in the larger (rhs) precision (or the combined safe dtype) the other option will do the cast and thus the operation in the lower precision.

You can avoid it

like this example Can use "unsafe" method via numpy.subtract

arr += image.flatten()

arr = np.add(arr, image.flatten(), out=arr, casting="unsafe")

But the fundamental problem is your numpy array have "int32" type value

Upvotes: 1

DataSciRookie
DataSciRookie

Reputation: 1200

for array you need the same type whenever you want to realise some operations as += or -= (additions or substracts)

Upvotes: 1

Related Questions