2 dimensions array of unknown size in Python

Question

I am pretty new at python and numpy, so sorry if the question is kind of obvious. Looking on the internet I could not find the right answer. I need to create a 2 dimensions (a.ndim -->2) array of unknown size in python. Is it possible? I have found a way for 1 dimension passing through a list but no luck with 2 dimension.

example

for i in range(0,Nsens):
    count=0
    for l in range (0,my_data.shape[0]):
        if my_data['Node_ID'][l]==sensors_name[i]:
            temp[count,i]=my_data['Temperature'][l]
            count=count+1
        else:
            count=count

Where temp is the array I need to initialize.

Very new myself. I think you will help yourself the most by looking at the NumPy module and how it handles multidimensional arrays. (Technicality: Python doesn't have multidimensional arrays. You can, however, have a list of lists.) — bob.sacamento
– bob.sacamento, Commented Nov 21, 2012 at 18:58
i would store data into dict of 1-d array first, having sensor_name as key. Then when you read all data i construct 2d array at that time you know how many sensors you have — yosukesabai
– yosukesabai, Commented Nov 21, 2012 at 18:58
from what I can see in your example maybe your code could be vecorized, you don't need a loop and you can get your resulting array by indexing. would be useful if you could give some example data — bmu
– bmu, Commented Nov 21, 2012 at 21:30

Alex I · Accepted Answer · 2012-11-22 08:30:10Z

3

This shows a fairly high-performance (although slower than initializing to exact size) way to fill in an array of unknown size in numpy:

data = numpy.zeros( (1, 1) )
N = 0
while True:
    row = ...
    if not row: break
    # assume every row has shape (K,)
    K = row.shape[0]
    if (N >= data.shape[0]):
        # over-expand: any ratio around 1.5-2 should produce good behavior
        data.resize( (N*2, K) )
    if (K >= data.shape[1]):
        # no need to over-expand: presumably less common
        data.resize( (N, K+1) )
    # add row to data
    data[N, 0:K] = row

# slice to size of actual data
data = data[:N, :]

Adapting to your case:

if count > temp.shape[0]:
    temp.resize( (max( temp.shape[0]*2, count+1 ), temp.shape[1]) )
if i > temp.shape[1]:
    temp.resize( (temp.shape[0], max(temp.shape[1]*2, i+1)) )
# now safe to use temp[count, i]

You may also want to keep track of the actual data sizes (max count, max i) and trim the array later.

answered Nov 22, 2012 at 8:30

Alex I

20.4k12 gold badges100 silver badges166 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Alex I Over a year ago

@user1842972: Does the code in the answer work for you? If yes, please remember to accept the answer.

Dr.EMG · Accepted Answer · 2012-12-13 23:56:48Z

1

Given your follow-up comment, it sounds like you are trying to do something like the following:

arr1 = { 'sensor1' : ' ', 'sensor2' : ' ', 'sensor_n' : ' ' }   #dictionary of sensors (a blank associative array)
                                                                #take not of the curly braces '{ }'
                                                                #inside the braces are key : value pairs
arr1['sensor1'] = 23
arr1['sensor2'] = 55
arr1['sensor_n'] = 125

print arr1

for k,v in arr1.iteritems():
    print k,v

for i in arr1:
    print arr1[i]

The Python Tutorials on Dictionaries should be able to give you the insight you are seeking.

answered Dec 13, 2012 at 23:56

Dr.EMG

15910 bronze badges

Comments

Community · Accepted Answer · 2017-05-23 11:43:57Z

0

in numpy you have to specify the size of the array while initializing. Later on you can expand the array if needed.

But remember expanding the array is not recommended and should be done as a last resort.

Dynamically expanding a scipy array

edited May 23, 2017 at 11:43

CommunityBot

11 silver badge

answered Nov 21, 2012 at 18:56

Harman

1,5811 gold badge19 silver badges31 bronze badges

Collectives™ on Stack Overflow

2 dimensions array of unknown size in Python

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related