[Scilab-users] problems with "mtlb_mode" variable when porting scilab 5.6 code to scilab 6.0

[christophk]

Hi there,

I used to program in a matlab environment and made a small script some time
ago that handles multiple files using "mtlb_dir". I am dealing with Excel
files and found out along the way that Scilab 5.6 couldn't handle to read
the new xlsx-format. (not sure whether that changed with the update to
scilab 6.0). This is why I ended up writing the following code example to
clear all xlsx files from the mtlb_dir list I originally created. The code
worked perfectly fine before but since I upgraded to Scilab 6.0 it started
to throw an error description of "Unknown field: dims".

Here is a code example to reproduce the error. The datapath actually needs
to contain one or more xls-files to end up throwing errors:

datapath = "C:\";
files = mtlb_dir(datapath + "*.xls");
exception = "xlsx";
j=0;
newfiles = [];
for i=1:size(files.name)
// == T if the current file is not xlsx
    if isempty(strindex(files(i).name,exception))
        j=j+1;
        newfiles(j) = files(i);
    end
end

I realized that "files" is of type struct(Mlist) whilst i create an empty
"newfiles" as a double.
That seems to be fine, however as long as I do a unindexed definition such
as
newfiles = files(i); I do not quite realise why that is the case.

So I tried to solve this problem by writing "newfiles = struct;" or
"newfiles = mlist;" for an empty file allocation instead of just "newfiles =
[]";. But that ends up giving me a variable type 130 instead of the type 17
I was aiming for.

Finally, I came up with a dirty workaround that I find very unsatisfying:
I've done an if-case exception.
So instead of writing
        newfiles(j) = files(i);
I made the following changes inside the code:
        if j == 1
            newfiles = files(i);
        else
            newfiles(j) = files(i);
        end

This works for fine for j == 1 but instead thows a different error as I
arrive at j = 2: "Undefined variable: mtlb_mode"

This is where I got stuck. I know I probably shouldn't use the "mtlb_dir"
function as Scilab is moving away from being a matlab clone. However, I
already built a huge code framework that is totally reliant on the structure
that is supplied by the mtlb_dir command. In other words, I'd rather go back
to Scilab 5.6 right now instead of having to go through hundreds of lines of
code to get the rest of my scripts working with another file handling
approach.

Please help!

Best regards,
Christoph



--
View this message in context: http://mailinglists.scilab.org/problems-with-mtlb-mode-variable-when-porting-scilab-5-6-code-to-scilab-6-0-tp4036751.html
Sent from the Scilab users - Mailing Lists Archives mailing list archive at Nabble.com.