About txnelson

txnelson

1. Please attach a sample data table so I can test the code against real data. 2. Please provide a mockup of what you expect the final data table to look like.

txnelson

To figure out what is going on here, all you need to do, is to determine the value of the variable k and it will tell you why referencing theTables[k] is pointing to an invalid subscript. Typically this is because you have either used a For() loop with k as the index value, and in the specification of the For() loop, you are incrementing k beyond the number of elements of theTables list. Or, the variable k what used in a previous part of the script, and when it is referenced at this point, the value of k is inappropriate. If you hover over any of the references to the variable k in your script, it will display what the current value of k is. You should be able to then find out where the value came from. one last note....since you are not initializing your variables at the beginning of your script, the values of the variables will be set to the values they were at the last time the script ended. Therefore, that can be causing the issue too.

txnelson

Both of the below pieces of code work great for me, as long as the deftb data table reference has been set, and the theTables list has been populated. Since the script did not work for you, what error messages were there in the log? MD = {"MD1", "MD2", "MD3", "MD4", "MD5", "MD6"}; For( k = 1, k <= N Items( theTables ), k++, deftb << New Column( MD[k + 1], numeric ) ); // or, why go through all of the work to create the list MD // when you can easily create the character string required // for the column name as shown below deftb = Current Data Table(); For( k = 1, k <= N Items( theTables ), k++, deftb << New Column( "MD" || char[k + 1]), numeric ) );

txnelson

I created a data table with a column called Place ID. I populated it with 100 random values that ranged from 1 to 850. I then ran the section of code that various Box and ActualLoc and it created a data table that looks like this: I do not see the intent with performing these operations. It appears to me(I could easily be wrong) that you do not have the understanding of some elements that you are using. Within your first For() loop, you create the same columns of Scenario, Box and ActualLoc 5 times. And since JMP will not allow a column have the same name, when you create Scenario, Box and ActualLoc during the first loop, those are the actual names it will use. However, on loop 2, JMP will force the names to Scenario 2, Box 2 and ActualLoc 2. Loops 3,4,and 5 repeat this behavior. The formulas used are also a little confusing. You are repeatedly using the specification of (:Place ID - i). Since this is being used in a formula, which means that the formula will be applied to each row in the data table, and the result of this calculation will be the value of the variable Place ID for the given row (i.e row 1 in my data table has the value 268) and it will subtract off the current value of i, which has the value of 1 on the first loop, 2 on the second loop, etc. Thus the for row 1, the values range from 267 to 263. I suspect you may be thinking that :Place ID -1 is referencing the column that comes before :Place ID. Can you help me out, and clarify what you are attempting to do to your data table with the column Place ID in it?

txnelson

I assumed that Box was referencing a memory variable called Box which is a list, and contains several elements, each one being the name of a variable. If this isn't it, what is Box?

txnelson

You can always add your own graphics to any of the JMP graphs and charts, using JSL. But that would not be the repositioning of the legend, but rather creating your own graphical display on top of the Graph Builder output. This capability is documented in the Scripting Guide in a section called "Add Scripts to Graphs"

txnelson

I am sure, that the JSL below, as written, will not work. The sample code that was provided had pointers to data tables not defined, so therefore, I am making some assumptions that may not be exactly correct. However, if you follow the logic and syntax I provide, you should be able to get an idea of how to have your one piece of code used to loop across whatever number of input parameters specified // By specifying the variable to hold the data table names // that will be created from the Subset with the By clause // the variable specified becomes a List with an element // for each table created theTables = dtparms << Subset( By( :Defecting_Level ), All rows, columns( :Parameter, :CL_LL, :CL_UL, :MD_LL, :MD_UL ) ); Show( theTables ); For( k = 1, k <= N Items( theTables ), k++, deftb << New Column( "MD1", numeric ); parameters = Column( theTables[k], "Parameter" ) << get values; deftb_colnames = deftb << get column names( string ); parameters = Associative Array( Eval( parameters ) ); deftb_colnames = Associative Array( Eval( deftb_colnames ) ); parameters << Intersect( deftb_colnames ); params = parameters << get keys; For( i = 1, i <= N Items( params ), i++, foundRows = deftb << get rows where( As Column( deftb, column( theTables[k],"Parameter")[i] ) < column( theTables[k], "CL_LL")[i] & Is Missing( deftb:MD1 ) == 1 ); deftb:MD1[foundRows] = column(theTables[k],"MD_LL")[i]; foundRows = deftb << get rows where( As Column( deftb, column( theTables[k],"Parameter")[i] ) < column( theTables[k], "CL_UL")[i] & Is Missing( deftb:MD1 ) == 1 ); deftb:MD1[foundRows] = column( theTables[k],"MD_UL")[i]; ); );

txnelson

// I am assuming that your syntax of // nrj j = dt1 << Summary( // was an attempt to create a pointer to each new data table // being created in the Loop, one for each column in the list // of columns called "Box" // To do that, one needs to create a new List variable, and then // populate the list by inserting the new data table reference // by using an "Insert Into()" on each loop // Create a list variable to hold the data table references mrj = {}; For( j = 1, j < 3, j++, newDT = dt1 << Summary( // Your reference to the list variable called "Box" // Group( Box(i) ), // incorrect for 2 reasons. 1. When referencing a // subscript, it is done using square brackets [ ], // 2. the index variable you are specifying is "i", // while the index variable for your For() loop is // a the variable "j" Group( Box[j] ), Freq( "None" ), Weight( "None" ), Link to original data table( 0 ), output table name( "Sum " || Char( j ) ) ); Insert Into( mrj, newDT ); );

txnelson

The documentation in the Scripting Index indicate that only 4 options are available

txnelson

Now you need to go to Help=>JMP Documentation Library and read both the Discovering JMP and the Using JMP documents!!! It will make your time with JMP far more enjoyable.

txnelson

I can assume that the issue falls outside of the code you provided. Here is a simple example, which when run against the Sample Data table called Big Class works fine in JMP 15 Graph Builder( Size( 528, 456 ), Show Control Panel( 0 ), Legend Position( "Inside Left" ), Variables( X( :weight ), Y( :height ), Overlay( :sex ) ), Elements( Points( X, Y, Legend( 12 ) ), Smoother( X, Y, Legend( 13 ) ) ), SendToReport( Dispatch( {}, "400", LegendBox, {Set Title( "" )} ) ) ); See if it works for you, and then check your complete code for possible issues

txnelson

OK....here is a display of a data table that has 200 rows, with 118 different IDs and contains a date and an Expense column. What I want is a data table with just one entry per ID and for the date to be the last date the person had a transaction, and I want to know the sum of the expenses. So I go to Tables=>Summary and it brings up the following Dialog window which I fill in the information to group by IDs, return the Maximum Date value and the Sum of the Expenses I then click on the OK button and it creates a data table with the summary statistics I could have also gotten the same results by using the Tabulate Platform Analyze=>Tabulate It is a drag and drop report writer, that once you have the report the way you want, you can go to the red triangle and request that a data table be made from the report

txnelson

To determine if a character column is actually a datetime value, is pretty straight forward, particularly if the structure is known. Then it is a simple mater to set an informat to the datetime format of the incoming columns, and then change the data type to numeric and the modeling type to coninuous.....oh yes, and the format to the date or datetime format of your choice.

txnelson

I assume this is the same question you subsequently added to a previous discussion topic. The answer is to use the Summary platform Tables=>Summary

txnelson

Use the Summary platform Tables=>Summary

txnelson

I don't understand what changing the values to integers would do to help you identify datetime values. Datetime columns already are the number of seconds since Jan, 1, 1904. Turning them into integers just means that you would just be eliminating the fractions of a second from the datetime value.

txnelson

try double clicking on the p-value and change the format to show more decimal places to see what the actual p-value is.

txnelson

Sounds like you need to contact JMP support support@jmp.com

txnelson

Start up JMP in the "Run as Administrator" mode. That should fix it from that point forward

txnelson

Here is one way of handling it Names Default To Here( 1 ); dt = Open( "$SAMPLE_DATA/big class.jmp" ); New Window( "big class - Graph Builder", vlb = V List Box() ); gb1 = Graph Builder( invisible, Size( 528, 456 ), Show Control Panel( 0 ), Variables( X( :weight ), Y( :height ) ), Elements( Points( X, Y, Legend( 1 ) ), Smoother( X, Y, Legend( 2 ) ) ), Where( :age == 12 ) ); (Report( gb1 ) << top parent)[Text Box( 1 )] << delete; vlb << append( Report( gb1 ) ); gb2 = Graph Builder( invisible, Variables( X( :weight ), Y( :height ) ), Show Control Panel( 0 ), Elements( Points( X, Y, Legend( 1 ) ), Smoother( X, Y, Legend( 2 ) ) ), Where( :age == 13 ) ); (Report( gb2 ) << top parent)[Text Box( 1 )] << delete; vlb << append( Report( gb2 ) ); gb3 = Graph Builder( invisible, Variables( X( :weight ), Y( :height ) ), Show Control Panel( 0 ), Elements( Points( X, Y, Legend( 1 ) ), Smoother( X, Y, Legend( 2 ) ) ), Where( :age == 14 ) ); (Report( gb2 ) << top parent)[Text Box( 1 )] << delete; vlb << append( Report( gb3 ) );

txnelson

Without seeing your full script, it is pretty hard to determine where the issue is. My other question is, why are you doing all of the deletes at a single point in your code, rather than having the delete code at the point where the graph is being produced? Or are you using a By clause in your code, which changes the Display Tree structure to a completely different hierarchy from the single stand alone graph.

txnelson

OK.....I appreciate the clearer definition. The script below seems to work on your test data Names Default To Here( 1 ); dt = Current Data Table(); dt << New Column( "Cumulative_Sum_Test_Qty" ); dt << New Column( "Cumulative_Sum_Fail_Qty" ); dt:Cumulative_Sum_Test_Qty[1] = dt:Test_Qty[1]; dt:Cumulative_Sum_Fail_Qty[1] = dt:Fail_Qty[1]; holdTest = dt:Test_Qty[1]; holdFail = dt:Fail_Qty[1]; For( theRow = 2, theRow <= N Rows( dt ), theRow++, If( dt:Cumulative_Sum_Test_Qty[theRow - 1] > 300, holdTest = :Test_Qty[theRow]; holdFail = :Fail_Qty[theRow]; curRow = theRow; While( holdTest < 300, curRow = curRow - 1; holdTest = holdTest + :Test_Qty[curRow]; holdFail = holdFail + :Fail_Qty[curRow]; ); dt:Cumulative_Sum_Test_Qty[theRow] = holdTest; dt:Cumulative_Sum_Fail_Qty[theRow] = holdFail; , dt:Cumulative_Sum_Test_Qty[theRow] = dt:Cumulative_Sum_Test_Qty[theRow - 1] + dt:Test_Qty[theRow]; dt:Cumulative_Sum_Fail_Qty[theRow] = dt:Cumulative_Sum_Fail_Qty[theRow - 1] + dt:Fail_Qty[theRow]; ) );

txnelson

Here is a simple function that will do the trick getNumber = Function( {input}, {defaultlocal}, numStart = Length( Word( 2, z, "0123456789" ) ); number = ""; While( Is Missing( Num( Substr( z, numStart, 1 ) ) ) == 0, number = number || Substr( z, numStart, 1 ); numStart++; ); number = Num( number ); ); value = getNumber( "wd*()578pqxkkl" ); Show( value );

txnelson

This should work gb = Graph Builder( Size( 531, 464 ), Show Control Panel( 0 ), Variables( X( :Step ), Y( :Name( "BR Metabolite/Count Data" ) ), Overlay( :Run ) ), Elements( Line( X, Y, Legend( 1 ) ), Points( X, Y, Legend( 2 ), Jitter Limit( 0.2018 ) ) ), Where( :Name( "Metabolite/Count" ) == "pH" ), SendToReport( Dispatch( {}, "Graph Builder", OutlineBox, {Set Title( "pH" ), Image Export Display( Normal )} ), Dispatch( {}, "graph title", TextEditBox, {Set Text( "pH vs. Step" )} ), Dispatch( {}, "Y title", TextEditBox, {Set Text( "pH" )} ) ) ); (Report( gb ) << top parent)[Text Box( 1 )] << delete;

txnelson

Here is a solution. I am assuming as soon as you go over 300 you restart. If you need to stop before you go over 300, the logic change is pretty simple. Names Default To Here( 1 ); dt = Current Data Table(); dt << New Column( "Cumulative_Sum_Test_Qty" ); dt << New Column( "Cumulative_Sum_Fail_Qty" ); dt:Cumulative_Sum_Test_Qty[1] = dt:Test_Qty[1]; dt:Cumulative_Sum_Fail_Qty[1] = dt:Fail_Qty[1]; holdTest = dt:Test_Qty[1]; holdFail = dt:Fail_Qty[1]; For( theRow = 2, theRow <= N Rows( dt ), theRow++, If( dt:Cumulative_Sum_Test_Qty[theRow - 1] > 300, dt:Cumulative_Sum_Test_Qty[theRow] = dt:Test_Qty[theRow]; dt:Cumulative_Sum_Fail_Qty[theRow] = dt:Fail_Qty[theRow]; , dt:Cumulative_Sum_Test_Qty[theRow] = dt:Cumulative_Sum_Test_Qty[theRow - 1] + dt:Test_Qty[theRow]; dt:Cumulative_Sum_Fail_Qty[theRow] = dt:Cumulative_Sum_Fail_Qty[theRow - 1] + dt:Fail_Qty[theRow]; ) );

txnelson

Here is the Select Where() clause you want to use dt << select where( :New Y <= 50000 | :New Y >= 50999 );

txnelson

Here is a different take on how to do this. It uses RSquare as an indicator of matching data. Names Default To Here( 1 ); dt = Current Data Table(); colNames = dt << get column names( continuous ); obj = dt << Response Screening( Y( Eval( colNames ) ), X( Eval( colNames ) ) ); dtCalc = obj << get PValues; dtCalc << select where(Round(:RSquare,4) != 1 | :X == :Y ); try( dtCAlc << delete rows ); dtCompare = dtCalc << subset( selected rows(0), columns("X","Y")); close(dtCalc,nosave); for each row( If(:X > :Y, hold=:X; :X = :Y; :Y = hold; ) ); dtFinal = dtCompare << Summary( Group( :X, :Y ), Freq( "None" ), Weight( "None" ), Link to original data table( 0 ) ); close( dtCompare, nosave );

txnelson

Here is a selection taken from the JMP Documentation Library, on the Statistical Details for Cpk The JMP Documentation Library is found under the Help pull down menu

txnelson

Here is how I would do it. There is an item I did in the JSL that were not required, but made the display look better. I used the Substitute function to make the Where clause more readable. Names Default To Here( 1 ); dt = Open( "$SAMPLE_DATA/Big Class.jmp" ); dt << set name( "Example" ); dt:age << set name( "B" ) << data type( character ); dt:sex << set name( "A" ); dt << delete columns( "name" ); Summarize( dt, aList = by( :A ) ); Summarize( dt, bList = by( :B ) ); NW = New Window( "Matrix OutPut", vlb = V List Box() ); For( i = 1, i <= N Items( aList ), i++, hlb = H List Box(); For( k = 1, k <= N Items( bList ), k++, If( N Rows( dt << get rows where( :A == alist[i] & :B == bList[k] ) ) > 0, hlb << append( V List Box( Eval( Substitute( Expr( biv = Bivariate( Y( :height ), X( :weight ), Where( :A == __aList__ & :B == __bList__ ) ) ), Expr( __aList__ ), aList[i], Expr( __bList__ ), bList[k] ) ) ) ) ) ); vlb << append( hlb ); );

txnelson

Removed at Discussion Author's request

Date Registered	‎06-22-2012 10:18 AM
Date Last Visited	12m ago
Total Messages Posted	4,886
Total Kudos Received	4391

Online Status	Offline
Date Last Visited	12m ago