Using specific row values for use in a formula with JSL

scottahindle

I would like to do something that I learnt a long time ago in VB, but am struggling with JMP scripting language. I am new to JMP.

I would like to run a For loop in a formula but use a cell array.

 

For example, I know I can reference a column as follows: j = Column ("a");

If I then say j[4] = 77 + 2;   I will get the value 79 in column “a”, row 4.

What, though, if I wanted to run a loop on Column "a", for example For(i=1,i<=20,i++, SOME FORMULA USING j[i]);

 

In the attachment I would like to use column-B Row [1] with column-C’s Rows 1 to 5. And then, column-B Row [6] with column-C’s Rows 6 to 10. My aim is to use the sequence in Column-A as an identifier, hence each time the sequence is of value 1 I hold this value for 5 subsequent uses of the formula where what goes into Column-D depends on the aritmetic performed on the contents of Columns B and C. (For example Column-D=Column-B-Value + Column-C-Value.) See the attachment to see in a picture what I mean.

 Any help would be great and much appreciated.

Example.jmp

[Untitled].pdf

txnelson

The following formula works for your attached example

targetrow = (Row() - :Name( "Column-A" )) + 1;
:Name( "Column-B" )[targetrow] + :Name( "Column-C" );

Remember, the formula will always loop through all of the rows, and the function, Row(), will always give you the row that is currently being worked on.  When using a formula in JMP, one needs to look at what results are needed when looking at a given row.  It is a different paradigm than VB

Here is another version, doing the For looping you are looking for

If( :Name( "Column-A" ) == 5,
	For( i = Row() - 4, i <= Row(), i++,
		:Name( "Column-D" )[i] = :Name( "Column-B" )[Row() - 4] + :Name( "Column-C" )[i]
	)
);
// The last value calculated will be the value for the current row, 
// so it needs to be handled as a special case
:Name( "Column-B" )[Row() - 4] + :Name( "Column-C" )[Row()];

I guess that technically, you could approximate a VB world, if one bypassed all processing for all of the rows, until the last row was detected, and then you could go back and loop through all of the rows under program control.....that seems like a lot of work

dt=current data table();
If(Row()==N Rows(dt),
<Do whatever processing you want>
);
<Process for the current(last) row>