topic Re: Create New Column with variables in Discussions

Create New Column with variables

fitri — Thu, 28 Nov 2013 06:17:19 GMT

Hi all,

I am trying to create a new column with formula, which represented by 2 variables as below:

dt << New Column("myNewColumn", Numeric, Continuous, Formula(:Ceiling( XXColumn / Size) * Size));

where:

XXColumn is one of column name in dt, selected in column dialog

Size = numeric value from column dialog

My attempt resulted a new column with formula with no value as formula create as below:

Ceiling(Column(XXColumn ) / Size) * Size

without actually assign it.

Any idea how I can accomplish this?

Re: Create New Column with variables

ms — Thu, 28 Nov 2013 07:26:46 GMT

Variables in formulas are not evaluated when created and should therefore be expressed literally. One way to circumvent that is to manipulate and parse strings.

Example:

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

cols = Column Dialog( Col1 = ColList( "X", Max Col( 1 ), Datatype( numeric ) ), Col2 = ColList( "Y", Max Col( 1 ), , Datatype( numeric ) ) );

XXColumn = cols["Col1"][1] << get name;

Size = cols["Col2"][1] << get name;

Eval( Parse( Eval Insert( "dt << New Column(\!"myNewColumn\!", Numeric, Continuous, Formula(Ceiling( ^XXColumn^ / ^Size^) * ^Size^))" ) ) );

Re: Create New Column with variables

fitri — Thu, 28 Nov 2013 07:45:33 GMT

Thank you MS.

The new column created with correct formula. But, the rows are empty. I need to go to the formula section on UI and click apply then only values prompts.

is it expected?

Re: Create New Column with variables

ms — Thu, 28 Nov 2013 08:25:22 GMT

No, it was not expected. For me (JMP 10 for Mac) it works without clicking "Apply". There was an inadvertant colon before "Ceiling..." that remained from your original code (I've now edited my code above). Maybe it helps to delete that.

Re: Create New Column with variables

fitri — Thu, 28 Nov 2013 09:05:57 GMT

mine is JMP8.0.0 Windows

with or without it still the same. T_T

Re: Create New Column with variables

ms — Thu, 28 Nov 2013 09:46:34 GMT

Try <<eval formula.

Eval( Parse( Eval Insert( "dt << New Column(\!"myNewColumn\!", Numeric, Continuous, Formula(Ceiling( ^XXColumn^ / ^Size^) * ^Size^), <<eval formula)" ) ) );

Re: Create New Column with variables

fitri — Thu, 28 Nov 2013 10:44:30 GMT

Its not working. The result still same.

In JSL, apply button == ? do you know?

Re: Create New Column with variables

ms — Thu, 28 Nov 2013 10:57:29 GMT

No. Eval Formula is supposed to be eqivalent as far s I know.

Try to separate the new column and set formula into separate expressions:

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

cols = Column Dialog(

Col1 = ColList( "X", Max Col( 1 ), Datatype( numeric ) ),

Col2 = ColList( "Y", Max Col( 1 ), Datatype( numeric ) )

);

XXColumn = cols["Col1"][1] << get name;

Size = cols["Col2"][1] << get name;

col = dt << New Column( "myNewColumn", Numeric, Continuous );

Eval( Parse( Eval Insert( "col << Set Formula(Ceiling( ^XXColumn^ / ^Size^) * ^Size^)" ) ) );

col << eval formula;

Re: Create New Column with variables

fitri — Thu, 28 Nov 2013 11:10:13 GMT

It did not work. Maybe I should explain what I am doing and share paste my code.

Step 1: Collect input from Column dialog

Step 2: Do Summary table with Min,Mean, Max for selected column + default column

Step 3: Create new column for Ceiling

Step 4: Do plotting using Graph builder

Code:

dt0 = current data table();

FN = dt0 << Get Name;

r = Column Dialog(

Bin = ColList( "X, Binning By", Max Col( 1 ) ),

Overlay = ColList( "Group by / Overlay by", Max Col( 1 ) ),

Split = ColList( "Y, Response",Max Col( 5 ) ),

Compare = ColList( "X-Axis Top", Max Col( 1 ) ),

HList( "Bin Size: ", BinSize = EditNumber( .1 ) ),

);

Bin = r["Bin"];

Overlay = r["Overlay"];

Split = r["Split"];

Compare = r["Compare"];

BinSize = r["BinSize"];

dts = dt0 << Summary(

Group(:Fixed, column(dt0,Overlay), column(dt0,Compare) ),

Mean( :Yield ),

Min ( column(dt0,Bin) ),

Mean( column(dt0,Bin) ),

Max ( column(dt0,Bin) ),

Min ( column(dt0,Split[1]) ),

Mean( column(dt0,Split[1]) ),

Max ( column(dt0,Split[1]) ),

Min ( column(dt0,Split[2]) ),

Mean( column(dt0,Split[2]) ),

Max ( column(dt0,Split[2]) ),

Min ( column(dt0,Split[3]) ),

Mean( column(dt0,Split[3]) ),

Max ( column(dt0,Split[3]) ),

Min ( column(dt0,Split[4]) ),

Mean( column(dt0,Split[4]) ),

Max ( column(dt0,Split[4]) ),

Min ( column(dt0,Split[5]) ),

Mean( column(dt0,Split[5]) ),

Max ( column(dt0,Split[5]) ),

statistics column name format( "column stat" ),

);

dts << Set Name (FN || " Summary");

n_Bin = Bin[1]||" Mean";

f_Bin = Column(dts, n_Bin);

col = dts << New Column("Binning", Numeric, Continuous);

Eval( Parse( Eval Insert( "col << Set Formula(Ceiling( ^f_Bin^ / ^BinSize^) * ^BinSize^)" )) );

Re: Create New Column with variables

ms — Thu, 28 Nov 2013 11:55:38 GMT

Try use n_Bin directly instead of f_Bin:

Eval( Parse( Eval Insert( "col << Set Formula(Ceiling( ^n_Bin^ / ^BinSize^) * ^BinSize^)" )) );

Re: Create New Column with variables

fitri — Fri, 29 Nov 2013 01:52:38 GMT

finally the magic happened !!!

here is the code i used..

col = dts << New Column("Binning", Numeric, Continuous);

Eval( Parse( Eval Insert( "col << Set Formula(:Ceiling( ^n_Bin^ / ^BinSize^) * ^BinSize^)<< eval formula" )) );

Thank You MS for the guidance; what is the diff on f_bin and n_bin i used?

Re: Create New Column with variables

ms — Fri, 29 Nov 2013 11:19:36 GMT

As far as I could tell from reading your code f_Bin refers to a column whereas n_Bin is a string, which is the data type that is safest to use here.

Re: Create New Column with variables

fitri — Mon, 02 Dec 2013 08:22:18 GMT

Thank you again, MS. I shall take note on this.