I have a table with more than 200 continues factor columns and few responses.
The columns are results of lab tests that look for the presence of substance in un object,
Each column is the amount of one substance.
These 200 columns can be divided in many ways (colors, purpose, activity, manufacture, etc.)
While analyzing the result I want to run all kind of tests and models on different groups out of the 200 columns.
What I'm looking for is a tool that allows me to easily select and un select different columns base on some deviation
What I think of is a sub table that the column name become rows.
The Sub table is connected to the main data table, and every row that I'm selecting in the sub table, automatically select the corresponded column in the main data table.
I can add columns that define the original columns (now rows). Then easily select different groups in them. Put them in and out of models or any other dialog box.
You may not need the extra table if you use something like this:
Names Default To Here( 1 ); // Get a table dt = Open("$SAMPLE_DATA/Big Class.jmp"); dtname = dt << getName; // Expression that is invoked whenever a new selection is made in the column selector selectionScript = Expr( colsSelected = fcs << getSelected; dt << clearColumnSelection; dt << selectColumns(colsSelected); ); // User interface New Window("Select Columns in "||dtname, fcs = Filter Col Selector( Data Table( dtname ), Width(300), Script(selectionScript) ); );
The 'red triangle' options in the filter might help you to find (then select) the columns you need, and it will also show any column groups you might have defined.
This 'red triangle' with its option exist in every dialog.
It can show me the group that I defined in the main table.
What I'm looking for is to be able to select in one click (or two) few columns that share some common attribute in a property (Not a property like "data type" or "has formula"), and then select other columns that share same Attribute in another property.
In my data I can divide my columns (which express level of presence of chemical substance in feathers) to few Groups.
I can divide them by influence on omen, or by Molecular formula, or by lethality, or by 'action techniques' and so on.
Each deviation creates a different group structure.
My question is if there is any tool in hand.
If not is there any simple "add in" that can be created.
The JSL above does select columns in the table, but perhaps was an incorrect interpretation of what you are asking.
When you say '. . . any tool in hand', do you mean as part of JMP, or as an add in that exists already? If so, I think the answer is 'no'.
Such an add-in could be written, but I think you would need to be more presise about the workflow that's needed to accomplish what you have in mind.
Do you already know JSL, or are you prepared to learn a little? Or do you have access to someone who could do it for you?
I'm aware of some solution that is possible:
If I open the 'Columns View Selector', select all relevant columns click on 'Show summery', then in the red triangle choose 'data table view' then the table created is connected to the main table and reflects my selection in row to the right columns in the underlined table.
I can than up date with all kind of columns which defines groups and work with it (Select matching cell, etc.), but the connectivity is disappears the moment I close the 'Columns View Selector'.
Is there any way to preserve that connectivity between both tables?