.ptp() not @njit compilable anymore?

[La-Li-Lu-Le-Loa]

• I have tried using the latest released version of Numba (most recent is
  visible in the change log (https://github.com/numba/numba/blob/master/CHANGE_LOG).
• I have included below a minimal working reproducer (if you are unsure how
  to write one see http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports).

I cant pass numpy arrays like the following to an njit compiled function that contains numpy.ptp().
Why is that ( I think here it says, that .ptp() is supported: https://numba.pydata.org/numba-doc/dev/reference/numpysupported.html)

a = [0.         ,0.         ,0.         ,0.         ,0.         ,0.
,0.         ,0.         ,0.00224828 ,0.00317533 ,0.00830591 ,0.00934349  
,0.80603907 ,0.80603907 ,0.80603907 ,0.80603907 ,0.80603907 ,0.94597275,
1.,        1.,        1.,        1.,        1.,        1., 
1.,       ]

import math
from numba import  prange
@nb.njit(nopython=True,parallel=True)
def anynan(array): 
   for i in prange(array.size):
       if math.isnan(array[i]):
           return True
   return False
@nb.njit
def scale_zero2one(x):
     
   x -= x.min()  
   x /= x.ptp()   
   if anynan(x):  
       x[np.isnan(x)] = 0               
   return x 

a = np.asarray(a,dtype=np.float32) 
print(scale_zero2one(a))

[gmarkall]

np.ptp is supported, but the .ptp method of array objects is not. Modifying your reproducer to define missing imports, define a, and use np.ptp instead results in a successful compilation:

import math
import numpy as np
from numba import prange, njit


@njit(nopython=True, parallel=True)
def anynan(array):
    for i in prange(array.size):
        if math.isnan(array[i]):
            return True
    return False


@njit
def scale_zero2one(x):

    x -= x.min()
    x /= np.ptp(x)
    if anynan(x):
        x[np.isnan(x)] = 0
    return x


a = [1, 2, 3]

a = np.asarray(a, dtype=np.float32)
print(scale_zero2one(a))

[stuartarchibald]

#6261 fixes