Transform content of cell array to strings

Hi,

I am importing some data, amongst other coordinate points (Lon/Lat), from a xlsx file using readtable. I then use table2array on some columns in order to get rid off the "table type" and be able to use the data for further analysis. The table2array transformation produces a 289x1 cell. Each cell contains only one line such as e.g.: 013° 39.541'E

This coordinate point I want to transform from degrees, minutes to a decimal coordinate using

LonHHMM = sscanf(CheckS, '%d %*s %f');
LonDeg = LonHHMM(1)+LonHHMM(2)/60;

The problem I am having is that I do not know how to extract the content of each cell as string. I do know that for a single cell I can use

CheckS = CPLon{1};

and then the above code to decimal format works fine. But how can I do this for all 289 cells? I tried using cellfun but I dont know what the first argument (@???) should look like. Or do I need to use a for loop in this case?

Your help is highly appreciated. Thanks in advance!

Oliver

Answer 1

[Thorsten]

cellfun is your friend here. The first argument of cellfun is an anonymous function of one argument, named c. cellfun then passes the content of each cell of your array to this function in the variable c.

(You could also replace c with CheckS, to match the naming in your post.)

X = cell2mat(cellfun(@(c) sscanf(c, '%d %*s %f'), CPLon, 'UniformOutput', false));
LonDeg = X(1,:) + X(2,:)/60

[Oliver]

Thank you very much for the swift response, it works like a charm! I managed to solve it with a for loop in the meanwhile and was quite surprised to see that the "for loop" solution runs faster than the cellfun code. Is this to be expected?

My code:

 LonHHMM = zeros(length(CPLon),2);
 LatHHMM = zeros(length(CPLon),2);
 for i = 1:length(CPLon)
    LonHHMM(i,:) = sscanf(CPLon{i}, '%d %*s %f');
    LatHHMM(i,:) = sscanf(CPLat{i}, '%d %*s %f');
 end
 CabLon = LonHHMM(:,1)+LonHHMM(:,2)/60;
 CabLat = LatHHMM(:,1)+LatHHMM(:,2)/60;

[Thorsten]

This is probably due to the fact that cellfun creates cells and requires the 'UniformOutput', false parameters, to work, although the output is uniform in this case, namely just two scalar values, which can be converted to a matrix with mat2cell. But this causes overhead which makes the cellfun solution slower.

But if you a big fan of cellfun you may still prefer the concise cellfun syntax, if the drop in performance is not too bad.

[Stephen23]

' "for loop" solution runs faster than the cellfun code'

cellfun can be much more convenient than a loop, and sometimes it is much tidier code too. But loops are often faster.

Answer 2

[Robert]

If all the lines of your data take the format of your example line, you could use

x = regexp(myCellArray,'([\d.]*)','match')
x = vertcat(x{:})
x = str2double(x)

to extract the numeric values of your strings.

Then, if your coordinates aren't all East, you will need to grab the direction and apply that to your data as required by your application. You could try

directions = cellfun(@(s)s(end),myCellArray)

If you are able to provide an example (small-ish) data set, we might be able to determine a cleaner way to load your data from the table.

[Oliver]

Hi Robert,

I attached a small data set from the xlsx file that I can share with you. Always looking for ways to improve my code :) Thanks!