cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Browse apps to extend the software in the new JMP Marketplace
Choose Language Hide Translation Bar

Selecting Cells based on color

Hello, 

After running the outlier analysis tool in JMP, I color coded the cells that contained outliers. How can I select all these cells and replace them with the name, "outlier"? The next step, would be to name all the other cells "normal". 

Let me know if this is feasible. A snapshot of the datasheet in appended in this post. 

MikeDereviankin_0-1607644587261.png

 

M. Dereviankin
4 REPLIES 4
txnelson
Super User

Re: Selecting Cells based on color

I am not aware of an easy way to get the cell color from a cell, let alone, select all of the rows for cells in a column, that has a specific color in it.  However, the way that I do that, is to extract the script for the column of interest, and then to search for the Color Cells area of the script, and parse out the rows.

I have attached a sample data table.  The data table was produced using the Outlier Analysis Tool. When the below script is run on that data table, it will find the cells in the column SS-E that have been colored light red(color 35 to JMP), and change the value for those cells to "outlier" and then change all of the other cells to "normal".

sc1.PNG

Please note, to change the cells to either "outlier" or "normal" the column had to be changed to a character column, and all of the original data are replaced.

Here is the script

names default to here(1);
dt=current data table();


x=char(:name("ss-e")<<get script);
x=substr(x,contains(x,"Color Cells"));

x=substr(x,contains(x,"{35, ")+5);
x=substr(x,1,contains(x,"}"));


myMatrix =matrix( parse(x));

dt:Name("ss-e") << data type(character);

dt << select rows(myMatrix);

dt:Name("ss-e")[dt<<get selected rows] = "outlier";
dt << invert row selection;
dt:Name("ss-e")[dt<<get selected rows] = "normal";
Jim
ian_jmp
Level X

Re: Selecting Cells based on color

As an alternative to the approach of @txnelson, it's maybe worth going back a step to think about how the table was made. The attached version of the 'Penta' sample data was made exploiting the 'Add To Missing Value Codes' option:

Screenshot 2020-12-11 at 12.03.00.png

You can then use code like this, which has the added benefit that you are not actually changing the data values in the table:

NamesDefaultToHere(1);

dt = DataTable("myPenta");

// Pick a column
myCol = Column(dt, "L4");

// Get the values deemed to be outliers
mvc = myCol << getProperty("Missing Value Codes");

// Delete this column property, else the subsequent 'getValues()' will return '.'
// rather than the cell value in the case of outliers
myCol << deleteProperty("Missing Value Codes");

// Get all the values
vals = myCol << getValues;

// Build the list to go into the 'Value Label' property
valList = {};
for(v=1, v<=NItems(vals), v++,
	if(Contains(mvc, vals[v]),
		// We have an outlier . . .
		InsertInto(valList, EvalExpr(Expr(vals[v]) = "Outlier")),
		// Else the point is 'Normal'
		InsertInto(valList, EvalExpr(Expr(vals[v]) = "Normal"))
		)
	);

// Add the 'Value Label' property
myCol << valueLabels(valList);

Re: Selecting Cells based on color

This works great! 

What is the syntax to iteratively loop through each column name instead of appending just one column name? How do you restrict this loop to only look through columns 1-10 for example? 

Appreciate the scripting help, this is great stuff. 

M. Dereviankin
hogi
Level XII

Re: Selecting Cells based on color

One can go via the script behind the columns to get all colored cells:
How to get the row number of colored cells in a column and assign it a tag (say "1") in a newly crea...