CODEWITHSUNDEEP
python-3.xmatlabdictionarystruct
ITA
ITA

Reputation: 3870

Python equivalent of array of structs from MATLAB

I am familiar with the struct construct from MATLAB, specifically array of structs. I am trying to do that with dictionary in Python. Say I have a initialized a dictionary:

samples = {"Name":"", "Group":"", "Timeseries":[],"GeneratedFeature":[]}

and I am provided with another dictionary called fileList whose keys are group names and each value is a tuples of file-paths. Each file path will generate one sample in samples by populating the Timeseries item. Further some processing will make GeneratedFeature. The name part will be determined by the filepath.

Since I don't know the contents of fileList a priori, in MATLAB if samples were a struct and fileList just a cell array:

fileList={{'Group A',{'filepath1','filepath2'}};{'Group B',{'filepath1', 'filepath2'}}}

I would just set a counter k=1 and run a for loop (with a different index) and do something like:

k=1;
for i=1:numel(fileList)
    samples(k).Group=fileList{i}{1};
    for j=1:numel(fileList{i}{2})
        samples(k).Name=makeNameFrom(fileList{1}{2}{j})
        .
        .
    end
    k=k+1
end

But I don't know how to do this in python. I know I can keep the two for loop approach with

for (group, samples) in fileList:
   for sample in samples:

But how to tell python that samples is allowed to be an array/list? Is there a more pythonic approach than doing for loop?

Upvotes: 1

Views: 263

Answers (1)

I. Amon
I. Amon

Reputation: 184

You could store your dictionary itself in a list and simply append new dictionaries in every iteration of the loop:

samplelist = []
samplelist.append(samples.copy()) % dictionary copy needed when duplicating

Accessing the elements in the list would then work as follows (For example the 'Name' field of the i-th sample):

samples_i_name = samplelist[i]["Name"]

A list of all names would be accessible by a simple list comprehension:

namelist = [samplelist[i]["Name"] for i in range(len(samplelist))]

Upvotes: 1

Related Questions