[Introduction]
I am trying to create a dictionary using tuple in python, but an error occurred and I cannot create it.
I'm sorry, but I'd appreciate it if you could tell me how to correct the error.
Thank you for your cooperation.
】What I want to realize したい
I am currently studying graph theory, and in the process, I am making the following dictionary.
dicts={(i_1,j_1):length_1,(i_2,j_2):length_2,...,(i_10,j_10):length_10}
i—Leaving node
j —Arriving node
length —The distance between i and j
Specifically, we aim to create the following dictionary by substituting the attached test.csv number.
dicts={(2,20):100, (4,18):200,...,(20,2):1000}
試What I tried and what I'm experiencing (error message) <
If you run the appropriate source code, you will get an error similar to the following:
I looked up the error online, but I couldn't understand it.
There seems to be a problem with numpy, but I really want to use numpy, so I haven't tried other methods of loading data (pandas, etc.).
TypeError:unhashable type: 'numpy.ndarray'
[Applicable source code]
With the aim of creating a dictionary, we have organized the following programs:
import numpy as np
data=np.loadtxt("test.csv",
delimiter=",",
skiprows = 1,
usecols=(1,2,3)
)
keys=[(data[i:i+1,0:1], data[i:i+1,1:2]) for i in range(0,10)]
values = [ data [ i:i + 1, 2:3 ] for i in range (0,10) ]
dicts = {}
for i in keys:
for x in values:
dicts[i] = x
print(dicts)
test.csv is as follows:
fidij length
1 2 20 100
2 4 18 200
3 6 16 300
4 8 14 400
5 10 12 500
6 12 10 600
7 14 8 700
8 16 6 800
9 18 4 900
10 20 2 1000
TypeError:unhashable type: 'numpy.ndarray'
This means that numpy.ndarray is unhashable and cannot be used as a dictionary key.
(Only hashable ones can be used as dictionary keys.)
You can do it like this.
Modified range(0,10) to range(len(data)).
import numpy as np
data=np.loadtxt("test.csv",
delimiter=",",
skiprows = 1,
usecols=(1,2,3)
)
# keys=[(int(data[i:i+1,0:1]), int(data[i:i+1,1:2])) for i in range(0,10)]
# values = [int(data[i:i+1,2:3]) for i in range(0,10)]
keys=[(int(data[i:i+1,0:1]), int(data[i:i+1,1:2])) for i in range(len(data))]#10 Deleted
values = [int(data[i:i+1,2:3]) for i in range(len(data))] #10 Deleted
dic = {n:values[i]for i, n in enumerate(keys)}
print(dic)# {(2,20): 100, (4,18): 200, ...}
You can do it like this
dct=dict(zip(map(tuple,data[:,:2],data[:,2])))
display(dct)
It looks like this when you break it down (corrected by comment: 1:3=>:2, 3=>2)
k=data[:,:2]#keys:numpy.ndarray
v=data[:,2]#values:array like
k_t=map(tuple,k)
dict(zip(k_t,v))
The numpy.loadtxt—NumPy v1.21 Manual has a keyword called unpack.
unpack:bool, optional
If True, the returned array is translated, so that arguments may be unpacked using x, y,z=loadtxt(...).When used with a structured data-type, arrays are returned for each field.Default is False.
You can also use this feature to write as follows:
import numpy as np
i,j,length=np.loadtxt(
'test.csv', delimiter=',', skiprows=1,
usecols=(1,2,3), dtype=int, unpack=True)
dic = dict(zip(zip(i,j),length))
print(dic)
#=>
{(2, 20): 100, (4, 18): 200, (6, 16): 300, (8, 14): 400, (10, 12): 500, (12, 10): 600, (14, 8): 700, (16, 6): 800, (18, 4): 900, (20, 2): 1000}
I understand that using pandas is easy to convert, so I will introduce it to you for your reference.
import pandas as pd
from print import print
df = pd.read_csv('test.csv')
dfx = df.set_index(['i', 'j']) ['length'].to_dict()
print(dfx)
# =>
{(2, 20): 100,
(4, 18): 200,
(6, 16): 300,
(8, 14): 400,
(10, 12): 500,
(12, 10): 600,
(14, 8): 700,
(16, 6): 800,
(18, 4): 900,
(20, 2): 1000}
© 2024 OneMinuteCode. All rights reserved.