Subset Table Keeps Showing Old Rows – How to Clear Memory Between Runs?

[#1]

I’m working on a JMP JSL script where I filter tables based on a user-provided list of values. I create filtered tables using Subset.

The issue I’m running into is:

1. On the first run, the filtered table shows the correct rows.

2. On the second run, if I input a new list, the filtered table sometimes still shows rows from the previous run, or combines old and new rows.

3. I’m using Names Default To Here(1) and Clear Symbols() at the end of the script, but that does not fully clear the table from JMP memory, so Subset seems to retain references to old rows.

Is there a recommended way in JMP to completely clear memory references for tables before creating new subsets?

hows to make Subset filtering repeatable across multiple runs without old data showing up?

[#2]

Can you provide example of your script? If you are using subset correctly, it won't show "old rows".

[#3]

could you please share a simplified version of your  code - as short as possible.
... starting with loading a data table , e.g.

dt = Open( "$SAMPLE_DATA/Big Class.jmp" );

Getting fast helpful answers to your questions 
Getting correct answers to correct questions quickly 

[#4]

Most of such timing issues can be fixed via wait(0).
Just add it after the subset command - and maybe another one before the subset command.

[#5]

Thanks. I missed pasting the code , here is the snippet for reference ; reading the input list values via a window and then following : 

// Directory path
pathDir = "J:\jmp-data\raw_data";
fileList = Files In Directory(pathDir);
Show(fileList);



selectedValues = {};
For(i = 1, i <= N Items(fileList), i++,
	For(j = 1, j <= N Items(targetSource), j++,
		If(Contains(fileList[i], targetSource[j]),
			Insert Into(selectedValues, targetSource[j]);
			Break();
		);
	);
);

countValues = N Items(selectedValues);
Show(selectedValues);

FilterByValueList = Function({sourceTbl, valueList, outName},
    tempTbl = sourceTbl << Duplicate Window();
    tempTbl << Clear Row States;

    For(i = 1, i <= N Items(valueList), i++,
        val = valueList[i];
        tempTbl << Row Selection(
            Select Where(:Batch Number == val, Current Selection("Extend"))
        );
    );

    resultTbl = tempTbl << Subset(
        Selected Rows(1),
        Output Table Name(outName),
        Invisible
    );

    Return(resultTbl);
);

// Example usage:

RunDataSubset = FilterByValueList(DataRuns, SelectedValues, "RunDataSubset");
MetaSubset    = FilterByValueList(MetaRuns, SelectedValues, "MetaSubset");

I also found the window that is being created to take user input, has global variables and that could be the reason for old data to show up: 

                    SelectedValuestext = Select_Values_test0 <<get text;
                    SelectedValuesText0 = {};
					i=1;

SelectedValuesText0 = {}; -  still has nothing so whenever I run the code again it should take new user input values if they are diff. 

[#6]

Where does the << Duplicate Window() come from and what is it supposed to do? 

JSL does have quite a lot of powerful functions you could use instead of using For() everywhere. You can go over lists faster using Filter Each, Transform Each, For Each and you can most likely do the selection without loop using Contains().

[#7]

I realised duplicate window doesn't exist, so I went ahead with creating the dt like this:

    dt = inputTable << Subset(All Rows, All Columns,Output Table("dt"),Invisible);

I do not know if there's an easier way. 
a = b   doesn't work. for data table , I was looking for copy table(), i found copy file(), 

Also creating this way requires the table to be closed everytime. I found that easier to deal and deliver. 

calcTable = Try( Data Table("KPI"), Empty() );

[#8]

I'm not sure why you are creating two subsets in this case

Names Default To Here(1); 

path = Convert File Path("$SAMPLE_DATA/", windows);
files = Files In Directory(path);

targetfiles = {"Big Class.jmp", "Probe.jmp"};

files_of_interest = Filter Each({target}, targetfiles,
	Contains(files, target);
);



create_filtered_table = function({dt, values, tablename}, {Default Local},
	rows_of_interest = dt << Get Rows Where(Contains(values, :Filename));
		
	dt_res = dt << Subset(Rows(rows_of_interest), Selected Columns(0), Output Table(tablename), Invisible);
	
	return(dt_res);
);


dt1 = New Table("table1",
	Add Rows(6),
	Compress File When Saved(1),
	New Column("Filename",
		Character,
		"Nominal",
		Compact(),
		Set Values({"Big Class.jmp", "a", "b", "c", "Big Class.jmp"})
	)
);

dt2 = New Table("table1",
	Add Rows(6),
	Compress File When Saved(1),
	New Column("Filename",
		Character,
		"Nominal",
		Compact(),
		Set Values({"Probe.jmp", "a", "b", "c", "Big Class.jmp"})
	)
);



dt_res1 = create_filtered_table(dt1, files_of_interest, "Table1");
dt_res2 = create_filtered_table(dt2, files_of_interest, "Table2");

[#9]

FilterByValueList = Function({sourceTbl, valueList, outName},
    tempTbl = sourceTbl << Duplicate Window();
    tempTbl << Clear Row States;

    For(i = 1, i <= N Items(valueList), i++,
        val = valueList[i];
        tempTbl << Row Selection(
            Select Where(:Batch Number == val, Current Selection("Extend"))
        );
    );

    resultTbl = tempTbl << Subset(
        Selected Rows(1),
        Output Table Name(outName),
        Invisible
    );

    Return(resultTbl);
);

// Example usage:

RunDataSubset = FilterByValueList(DataRuns, SelectedValues, "RunDataSubset");
MetaSubset    = FilterByValueList(MetaRuns, SelectedValues, "MetaSubset");
//SelectedValues is a list