CODEWITHSUNDEEP
python-3.xsignaturenumbanamedtuple
NNN
NNN

Reputation: 593

Numba namedtuple signature

I'm trying to specify the return type for a namedtuple in Numba and I am not able to do so. Could someone help? Consider the following minimal code:

import numba as nb
from   collections import namedtuple

NT = namedtuple('NT',['sum','sum2'])

@nb.njit((nb.types.NamedTuple([nb.float64,nb.float64],NT))(nb.int64,nb.float64[:,:]),fastmath=True)
def arrsum_njit(nn,xx):
    arraysum = 0.0
    out = NT(sum=arraysum,sum2=arraysum)
    return out

I get the error

No conversion from NT(float64 x 2) to NT(float64, float64) for '$20return_value.7', defined at None

File "numbanamedtuple.py", line 10:
def arrsum_njit(nn,xx):
    <source elided>
    out = NT(sum=arraysum,sum2=arraysum)
    return out
    ^

During: typing of assignment at numbanamedtuple.py (10)

File "numbanamedtuple.py", line 10:
def arrsum_njit(nn,xx):
    <source elided>
    out = NT(sum=arraysum,sum2=arraysum)
    return out

Upvotes: 3

Views: 1875

Answers (3)

Michael
Michael

Reputation: 7756

Here is the full example using NamedUniTuple:

import numba as nb
import numpy as np
from   collections import namedtuple

NT = namedtuple('NT',['sum','sum2'])

@nb.njit(nb.types.NamedUniTuple(nb.float64, 2, NT)(nb.float64,nb.float64[:,:]))
def arrsum_njit(nn,xx):
    arraysum = 0.0
    out = NT(sum=arraysum,sum2=arraysum)
    return out

arrsum_njit(1, np.array([[1.], [2.]]))
# >>> NT(sum=0.0, sum2=0.0)

Upvotes: 1

VovaM
VovaM

Reputation: 390

The problem is "overoptimized" numba compiler (bug). Add a variable of a different type to the tuple to tell the compiler to use a heterogeneous tuple (internal class).

import numba as nb
import numpy as np
from collections import namedtuple

NT = namedtuple('NT',['sum','sum2','dummy'])

@nb.njit((nb.types.NamedTuple([nb.float64,nb.float64,nb.int64],NT))(nb.int64,nb.float64[:,:]))
def arrsum_njit(nn,xx):
    arraysum = 0.0
    out = NT(sum=arraysum,sum2=arraysum,dummy=1)
    return out

arrsum_njit(1, np.array([[1.], [2.]]))
# >>> NT(sum=0.0, sum2=0.0, dummy=1)

Upvotes: 3

Golden appendages
Golden appendages

Reputation: 116

Use NamedUniTuple instead. It's a numba spec type for homogeneous namedtuples.

Upvotes: 2

Related Questions