Try the Materials Informatics Toolkit, which is designed to easily handle SMILES data. This and other helpful add-ins are available in the JMP® Marketplace
Created:
Jan 27, 2023 07:15 AM
| Last Modified: Jun 8, 2023 9:40 AM(1225 views)
I just noticed that Graph Builder/BoxPlot has some problems with columns of modeling type Multiple Response.
this gets apparent if non-multiple-response entries are used. In plot 1, a male subset is showing up in the female group on the left.
This is not the case if color is used instead of Overlay (second plot). The wrong female plot looks like the 3rd plot - where on the x axis no distinction is made between male and female.
Just curious...if you think you found a bug, you should report this to support@jmp.com I'm not sure why you report it here on the forum. Unless you want us to confirm what you are finding, then it would be useful to attach the JMP table you are using.
"All models are wrong, some are useful" G.E.P. Box
Actually, I cannot imagine what could be wrong with the code.
But sometimes, it's a small comma or point.
Sure: If somebody confirms that this behavior is not like expected, I will send a link to this post to support@jmp.com.
Thank you for the comment about the data table. At first I didn't get the actual "message" - it was quite implicitly. Direct answer to the question: the data table is from the sample directory - so everybody should have the file on his computer. But you are right - no issues with this file --> when posting the script, I missed the line with the modeling type conversion.
I just added it now. So, you should be able now to reproduce the described behavior with the file from your sample directory.