Difficulty with the Distance Matrix Reader module

Hi all,

   I'm trying to read in an external distance matrix for use with a workflow including K-Medoids clustering and MDS, but am having some trouble. The distance matrix was calculated in PAST (I need Hamming's distance because of my data type), and was copied into a text file so that it is a symmetric 39 x 39 matrix of floating point numbers (diagonals all 0.0), and with row and column headers for each. Every cell is separated by a ";". I believe I am selecting all the proper checkboxes (I've actually tried all possible combinations to be sure), but I get the following error:

ERROR     Distance Matrix Reader     Execute failed: ("ArrayIndexOutOfBoundsException"): 40

    It seems that there may be some problem at the very end of the file? I'm attaching the text file containing the distance matrix to this post... Any ideas for how to make this work would be greatly appreciated!

FYI, I was able to get this working by pasting the matrix into Libreoffice Calc instead of directly into Gedit. Calc allows you to specify the encoding of the text file (I chose Unicode), which I think may have been the problem. Hope this helps other people who ecounter this issue! :)

Hi isaacullah

I was just looking into your problem, but you found the solution faster.

We will try to solve the problem.

Cheers, Iris

Your file contains an empty line at the end which confuses the reader. The error message offers room for improvement, though.

Thanks for the extra help! I thought it was the extra line, but I did try to delete it in a normal text editor (gedit), and it still had issues reading the file. It may have put out a slightly different error after trying that, however, but I'm not sure now if that is truely the case. What I'm getting at is that the extra line is certainly a problem, but I think I may have been having more than one problem! I suspect that the character encoding may also have been a problem. This might be because I'm using KNIME on linux (xubuntu 12.04), but generated the distance matrix with PAST running under wine (windows emulator). So, if I just copy and paste the matrix into gedit and save to an ASCII file with a .csv extention, I get conflicted character encoding (some windows, some linux), but if I put it into Libreoffice and save as a "real" .csv, it saves it exactly as the character encoding I tell it in the dialog that pops up... Hope that helps!


Oh, and yes, the error message could be more helpful! :)