converting string to num and table
Show older comments
I have a table containing string data that represents cyclone certain and uncertain tracks. Each cell in the table may contain multiple track numbers (seperated by a space), where the presence of 'x' beside a number indicates uncertainty in that track. For example, in tempData(7,1) = "30 25x", '30' is a certain track, and '25' is uncertain.
I need to write a code that parses each cell in a column of tempData, separates the track numbers and their uncertainty markers ('x'), and outputs a (in this case) 4-column table for each column in tempData. The headers of the output table should be track1, uncert1, track2, uncert2, corresponding to the parsed tracks and their certainty indicators (marked by 0 or 1). If the input has 3 tracks, the then output table should have a 6 column table
Finally, I want to store each output table in a cell array marked by the varname of the input column (y98, y99, y20, etc), as I need to process over 200 columns in my original data.
I have attached a sample input (tempData) and output for the first coumn that i want (tempDataOut).
Thanks in advance.
Accepted Answer
Tip for the future: upload your raw data, not a data extract that you have already imported and modified.
S = load('sample1.mat')
S.tempData % lets take a look at the input data
That table data is a bit of a mess, so the first thing we will do is convert the table data to a consistent type (fixing the data file import would make this task easier/superfluous):
V = S.tempData.Properties.VariableNames;
T = varfun(@mystring,S.tempData)
C = varfun(@myvarfun,T, 'OutputFormat','cell');
C = cellfun(@myrename,C,V, 'uni',0)
S.tempDataOut % your expected output for comparison against C{1}
C{:} % all output tables
function S = mystring(C)
% Convert cell array of mixed numeric/char/string to string array.
C(cellfun(@isempty,C)) = {''};
S = string(C);
end
function T = myvarfun(S)
% Convert string array to table (with certain & uncertain columns).
assert(~any(contains(S,'-')),'Minus character is not supported')
T = table(S);
C = regexprep(S,'(\d+)x','-$1');
N = numel(C);
M = nan(N,0);
for k = 1:N
V = sscanf(C{k},'%f');
M(k,1:numel(V)) = V;
end
U = +(M<0);
M = abs(M);
for k = 1:size(M,2)
T = addvars(T,M(:,k),U(:,k),'NewVariableNames',["track","uncert"]+k);
end
end
function T = myrename(T,v1)
% Rename first table column to v1.
T.Properties.VariableNames{1} = v1;
end
3 Comments
Peter Perkins
on 22 Aug 2024
This is a nice bit work, Stephen23. I woulda used strings instead of cellstrs from the getgo, but other than that ...
Sarvesh
on 23 Aug 2024
Stephen23
on 23 Aug 2024
@Peter Perkins: done.
More Answers (1)
Karanjot
on 22 Aug 2024
Hi Sarvesh,
To achieve this task, you can write a script in MATLAB that processes each column of your tempData table, parses the track numbers and their uncertainty markers, and stores the results in a cell array. Here's a step-by-step guide to help you accomplish this:
- Initialize Variables: Create a cell array to store the output tables, and define the headers for the output tables.
- Parse Each Column: Loop through each column in tempData, and for each cell, split the string by spaces to separate track numbers and uncertainty markers.
- To carry out the string manipulation as described above, You can use the strsplit function to split a string into parts based on a delimiter and the endsWith function to check if a string ends with a specified substring. For example:
parts = strsplit('30 25x', ' '); % Splits into {'30', '25x'}
isUncertain = endsWith('25x', 'x'); % Returns true
- Create Output Table: For each parsed cell, determine the number of tracks and uncertainty indicators, and create a table with appropriate headers.
- Store Results: Store each output table in the cell array with the corresponding variable name.
To know more about the functions mentioned above, Please refer to the following documentation:
I hope this helps!
Categories
Find more on String in Help Center and File Exchange
