lsline error message and how to deal with it

I try to use lsline in scatter plot. The scattor plot works fine. But for lsline I got an error message

"Warning: No allowed line types or scatterplots found. Nothing done."

Please advise.

Answer 1

[Star Strider]

With only the error message, it is not possible to determine exactly what the problem is.

From the lsline documentation:

‘lsline ignores data points that are connected with solid, dashed, or dash-dot lines ('-', '--', or '.-') because it does not consider them to be scatter plots. To produce scatter plots, use the MATLAB® scatter and plot functions.’

So if the plot appears to be correct, check the LineStyle you plotted them with. Change it if necessary to be compatible with lsline.

[alpedhuez]

Thank you. I think it is because some values are -inf so I need to remove them.

[alpedhuez]

I tried

https://www.mathworks.com/matlabcentral/answers/163545-remove-inf-and-nan-values-in-a-point-cloud-in-the-fastest-way

but told that this does not work for tables or timetables. How can one deal with this issue for table and timetable?

[Star Strider]

Doing logical operations on a table object (for example, to find Inf values in order to convert them to NaN values) is not easy, and apparently cannot be vectorised as would be straightforward for an array. (At least that is not in the documentation that I can find.) Probably the easiest way is to convert them to arrays, then do the logical tests, substitutions, and then use fillmissing to eliminate the NaN values. The problem with this is that the variable and row names (if they exist in the original table) are then lost.

A less efficient way that retains the table information is for example:

M = 1./randi([0 10], 7);                                % Create Matrix With ‘Inf’ Values
T1 = array2table(M);                                    % Convert To Table
T2 = T1;                                                % Copy
L2 = varfun(@isinf,T2);                                 % Find ‘Inf’ Values
[r,c] = find(table2array(L2));                          % Logical Array
for k = 1:numel(r)
    T2{r(k),c(k)} = NaN;                                % Loop To Replace Individual ‘Inf’ Values With ‘NaN’
end
T2 = fillmissing(T2,'linear');                          % Fill ‘NaN’ Values (Choose Appropriate ‘method’)

In spite of its inefficiencies, it eliminates the Inf values and fills them with interpolated data.

It would be a significant improvement to table objects if it was easy to vectorise these replacements, or have fillmissing optionally accept Inf as a missing value (with a Name-Value pair argument, if necessary).

.

[alpedhuez]

https://www.mathworks.com/matlabcentral/answers/561824-how-to-remove-a-row-that-contains-inf-from-a-table

[Star Strider]

That removes the entire column — not row — that contains Inf values.

It does not change the individual Inf values. My code — for what it’s worth — changes the individual Inf values and replaces them with interpolated (or, if you choose, constant) values.