np.argsort cannot work for matrix more than 1D

[isaac-you]

import numpy as np
from numba import jit 

@jit(nopython=True)
def f1(m1):
    return np.argsort(np.argsort(m1)) + 1.0

m1 = np.random.randint(10,size=(5,3))

f1(m1)

TypingError: Failed in nopython mode pipeline (step: nopython frontend)
Invalid use of Function(<function argsort at 0x7f5abc2cce18>) with argument(s) of type(s): (array(int64, 2d, C))
 * parameterized
In definition 0:
    All templates rejected with literals.
In definition 1:
    All templates rejected without literals.
This error is usually caused by passing an argument of a type that is unsupported by the named function.
[1] During: resolving callee type: Function(<function argsort at 0x7f5abc2cce18>)
[2] During: typing of call at <ipython-input-4-0ce306685c59> (3)


File "<ipython-input-4-0ce306685c59>", line 3:
def f1(m1):
    return np.argsort(np.argsort(m1)) + 1.0
    ^

This is not usually a problem with Numba itself but instead often caused by
the use of unsupported features or an issue in resolving types.

np.version:'1.17.2'
numba.version:'0.45.1'

[esc]

@isaac-you thanks for reporting this. I can confirm that this doesn't work and have converted the issue into a feature-request.

[EPronovost]

I've started working on this. I'm most of the way done of an implementation that works for IntegerLiteral axis arguments. From there I plan to extend it to support a limited range of Integer arguments (similar to np.sum). I'll put up a PR within a few days.

[esc]

@EPronovost 👍

[ucasiggcas]

hi,could you help me ?
#6285

[stuartarchibald]

Closing as duplicate of #4479.