About nikles

nikles

Thanks Jim - this is exactly what I was hoping for. One further question though. If no value ordering is detected, I plan to go ahead and set the column property within the script using just an empty list {}. It will still be up to the user to order the values though in the Column Info box. So once I have it open, is there a way to highlight the value ordering property? That is, instead of opening the dialog as it is in the first image below, I open it as it is in the second image: Not crucial, but would be nice to have if it were possible. Perhaps an undocumented option to the Main Menu command, or even an xpath command to act on the Column Info window? -Stefan

nikles

Hi. I'd like to know if there's a way to open the column info dialog within a script. For example, I have a script that requires the columns to have the value ordering property enabled. The script checks the columns for this property, and if not present, it creates an error window that tells the user to go back and fix it themselves. It would be nice to bring up the window for them though, preferably with the value ordering property already selected. I'm hoping for something similar to the dt << Recode message, which opens up the Recode dialog. However I can find no similar message for column info (e.g. dt << ColumnInfo). Please let me know if any ideas. Thanks.

nikles

Thanks @ih . That seems to have worked well. For some reason I get an error when I try to accept your answer as the solution - but this was the solution.

nikles

Update - Optimize Display does not do exactly what I'd like. When I run this command, it makes an effort, but it still does not make the larger columns wide enough. I've observed cases where the column is still too narrow to display the entire value in the column (truncated with a "..."), or the column heading has to be wrapped around to a second line.

nikles

Hi. I'd like to know if there is a good way to get the number of frame boxes in a graphbuilder plot. I have a GB plot that contains 104 FB (8 X-groups, 13 Y-groups). I've written a script to change the marker size for each FB. Summarize(dtstack, frame_lis = By(:Freq, :Image, :Region)); For(i=1, i<=NItems(frame_lis[1]), i++, Try( Report(gb) << Dispatch({}, "Graph Builder", FrameBox(i), {Marker Size(1)} ) ) ); (credit to @ms in https://community.jmp.com/t5/Discussions/Can-t-change-all-marker-size-in-Graph-Builder-by-script/m-p/3928 ). But in my case, the Summarize statement overestimates the number of FB, as I've excluded several rows in my table containing certain values of Region, Image, or Freq. This script works, but once the counter i exceeds the the actual number of FB in the GB plot, I begin getting error messages in the log ("Cannot find FrameBox[ "Graph Builder" ] at {}"). The "Try" statement was an attempt to prevent the error messages, but did not work. At any rate, I'd prefer to avoid this by just computing the correct number of frame boxes to begin with. Anyone have any ideas?

nikles · ‎09-24-2020

Hi. I'm having problems re-ordering my add-ins in the Add-Ins menu. I have several add-ins and I want to list them alphabetically to make locating them easier. To order the add-ins, I've been going to the menu bar, View > Add-Ins. I select the add-in I wish to move, and reorder it by changing the Load Order. This moves it up/down in the list. After, I check the Add-Ins menu, and confirm it is in there correct order: But once I exit JMP and restart, the add-in is moved back to the end of the list. I've tried unregistering/re-installing the add-ins with no luck. I've also tried deleting the entire add-in from the Add-Ins folder and creating a new version. Still won't work. There doesn't seem to be anything special about this particular add-in to make it different from the others. What am I missing? I'm using JMP15.2 Pro on a Mac with Catalina OS. Thanks.

nikles · ‎09-23-2020

Added to the Wish List: https://community.jmp.com/t5/JMP-Wish-List/Operator-to-set-all-column-names-using-a-list-of-strings/idi-p/311606

nikles · ‎09-23-2020

Hi. I'm looking for a scripting method to change the column names in my table to names I have contained in a list. I'm thinking something like: dt << Set Column Names(newcolname_lis)? This command would change the name of the ith column in the table to the ith item in newcolname_lis, starting at column 1. The reason for this request: I have tables that can contain >20k columns. I need to change the column names to merge with the rest of my data tables. Using a For loop to change the names takes several minutes. Ideally this could be done using a command in a few seconds. This is a follow up to a question I posted earlier: https://community.jmp.com/t5/Discussions/Set-column-names-more-efficiently/m-p/311439#M56450 Thanks.

nikles · ‎09-23-2020

Thanks for the advice Jim. At least I have some confidence that the way I'm currently doing it is the fastest. Will add to the wish list.

nikles · ‎09-23-2020

Thanks! Not a crazy idea at all. I tried this as well, but creating the table is equally slow.

nikles · ‎09-22-2020

Hi. I'm looking for a scripting method to efficiently change the column names in my table to names I have contained in a list, without using a for loop. Is there an operator or function that can perform this? I'm thinking something like: dt << Set Column Names(newcolname_lis)? With a for loop, I can do it this way: ... col_lis = dt << Get Column Names("String"); For(i=1, i<=NItems(col_lis), i++, dt:i << Set Name(newname_lis[i]); ); But, my tables can contain > 20k columns, and this takes several minutes. Any ideas on speeding this up? Thanks.

nikles · ‎08-18-2020

Thanks Jeff. Exactly what I'm looking for.

nikles · ‎08-18-2020

Hi. Suppose I have a string "cat,dog,,,,mouse,". Is there a simple way to break this string into a list of substrings in such a way that the result is {"cat", "dog", "", "", "", "mouse", ""}? I've been using the Words() function, but it seems to treat consecutive delimiters in series as just one: //Desired list result should be {"cat", "dog", "", "", "", "","mouse", ""} Names Default to Here(1); str = "cat,dog,,,,mouse,"; lis = Words(str, ","); //This returns {"cat", "dog", "mouse"} I tried to format the string ahead of time using Substitute: newstr = Substitute(str, ",,", ", ,"); //Returns "cat,dog, ,, ,mouse," lis = Words(newstr, ","); //Returns {"cat", "dog", " ", " ", "mouse"}. finalis = Substitute(lis, " ", ""); //Returns {"cat", "dog", "", "", "mouse"} Note that there are still 2 consecutive delimiters between dog and mouse after the first Substitute() command. Thus, following the Words() function, one of the empty items between dog and mouse is still missing. Also note that the empty item at the end of the string is missing from the list. Does anyone know of a way to do this, preferably one that does not involve a for loop?

nikles · ‎06-29-2020

I would like a command to return the pathname of the JSL script. For example, if I run a script "/Users/John_Doe/Documents/JMP/JMP Scripts/myscript.jsl", and I would like to have a command like this inside the script: thispath = Get Current Script Path(); Show(thispath); And I would like this to return: "/Users/John_Doe/Documents/JMP/JMP Scripts/myscript.jsl" The reason for this request: I have a "common functions.jsl" script that many of my scripts call using "Include". It contains many frequently used user functions that I don't want to copy into each new script I write. I also like to share my scripts with others in my group, along with "common functions.jsl". I can require that my group members place common functions.jsl in the same place relative to the JSL calling it (e.g. the same folder, or parent folder), but I can't control what the overall path names will be. If I had the path name of the calling script, I could easily use that to figure out where "common function.jsl" is (e.g. "/Users/John_Doe/Documents/JMP/JMP Scripts/common function.jsl"). I cannot use Get Default Directory(). I've found this command returns only the path of the most recent file opened using the Open File dialog. This works if the user opens the script and immediately runs it. The problem is if the user opens another script after opening "myscript.jsl", but then tries to run "myscript.jsl". It this case Get Default Directory() will return the path of the other script, not "myscript.jsl". Also, Get Default Directory() does not get updated if a script is opened from the Recent Files list. Any suggestions, or is this something that doesn't exist yet? Thanks for your help.

nikles · ‎06-29-2020

Thanks @gzmorgan0. It took a little time for me to grasp, but this is a nice solution. I appreciate the help!

nikles · ‎06-29-2020

Thanks for the suggestion ih. I looked into this, but regret to say I'm not really knowledgeable enough on the Run Program command to make this work. At first glance, it appears to require an actual file in the "Executable" field, rather than just a string. This strays into unknown territory for me unfortunately. I've created a wish list item that is related to this: Possible to have an add-in search for a script and run it? I appreciated the questions at the end of your reply, and I tried to address those questions in this wishlist item post. Thanks again for you help.

nikles · ‎06-29-2020

I would like a way to have an add-in do the following: 1. Find a given jsl script on the user's computer 2. Run that jsl. The reason for this request: When I develop scripts for my group, I would like for them to use Add-ins rather than scripts. I've found this increases ease-of-use and adoption rate. However, each time I update to a script used in an Add-in: a. I need save the updated script, then re-save it to the Add-in (Using the "Run JSL in this file:" option. Note that this saves a static copy of the script to the Add-in. Subsequent changes to the JSL are not reflected in the Add-in). b. Push the Add-in to our git repo. c. Notify all the users that a new version of the add-in has been created, and they will need to re-install the add-in in order to use. I'd like to not have to bother my group with constant update notifications that require their action. Instead, I'd like to make my add-in a simple "Include" command that points to the jsl file on their computer (Using the "Run this JSL:" option instead). The jsl file will be updated whenever they pull from the repo, and the add-in installed on their computer simply call that each time. Thus, the add-in would not require any changes on the users' computers. The problem is that the "Include" command in the add-in will somehow have to know where to find the jsl on the user's computer. I can't assume that every user will have the same directory structure on their computers, and can't make that a requirement either. It would be nice if the add-in had a feature where it could be given just a *.jsl file name and it would search the user's computer for it. Even nicer would be if it could save this location somewhere, so that subsequent calls of the add-in could just use this location and wouldn't need to search the entire computer. Possibly a slightly easier-to-implement option would be to have the Add-In open a file search window if it can't find the specified jsl file in the default search path, and let the user navigate to the name. Again though, I would want the addin to save this location somewhere so that the user would only need to do this once. Hope this makes sense. Thanks!

nikles · ‎06-25-2020

Hi. Is there a way to have a script search the user's computer for a given filename and return the path? Im using a Mac and JMP14.3 atm. I'm interested in creating an add-in that calls another script on the user's computer, but I won't know where the user has stored the script. In the add-in, I'm thinking something like this: //Add-in code path = [command or function to search for "myscript.jsl" on user's computer and return a path]; Include(path, New Context); //Things I've explored: //Not certain what the "search" option does, but seemed worth a go. //This only returns my default directory with "myscript.jsl" appended //(even though there is no such file in that location) path = ConvertFilePath("myscript.jsl", search); //There's also this, but same result as above: Set File Search Path("$HOME/"); path = ConvertFilePath("myscript.jsl", search); //As a last resort, I looked at this, thinking maybe if I could get a list of all files, //I could then search the list for myscript.jsl. Unfortunately I cannot get the "recursive" //option to work on my machine (Mac, JMP14.3). I imagine this would be a very slow method //though as there are 1000s of files in a typical user's home directory. y = Files In Directory("$HOME", "recursive"); //Returns {} y = Files In Directory("$HOME", "Recursive"); //Returns {} y = Files In Directory("$HOME", recursive); //Returns {} y = Files In Directory("$HOME", Recursive); //Returns {} y = Files In Directory("$HOME", Recursive(1)); //Returns {} //All options above just return {}. I'm curious if this is unique to my setup (Mac, JMP14.3). Other than that, I'm out of ideas. Does anyone have any suggestions? Thanks.

nikles · ‎06-19-2020

Hi. I'm interested in knowing there is a scripting method to read the displayed values for an axis in graphbuilder and return them as a list. IOW, is there a function or a message that can "get" the axis values from a graphbuilder reference or an axis box? Something like: xval_lis = gb << Get X Axis Values. For instance, suppose I have the following graph: dt = Open("$SAMPLE_DATA/Big Class.jmp"); gb = Graph Builder( Size(513, 406), Show Control Panel(0), Back Color("White"), Variables(X(:age), X(:sex, Position(1)), Y(:weight)), Elements(Points(X(1), Y, Legend(1)), Smoother(X(1), Y, Legend(2))), Local Data Filter( Add Filter( columns(:age), Where(:age == {12, 13, 14}), Display(:age, Size(170, 96), List Display) ) ), SendToReport(Dispatch({}, "", AxisBox, {Select})) ) Here, since the user has excluded certain ages using the local data filter, not all of the ages in the table are shown in x-axis. In this case I would like a way to read off which ages are actually displayed in the graph. This question applies only when a nominal or ordinal-type variable is used for the axis (no continuous variables). I've considered the following scripting method: 1. Read the local data filter 2. Create a subset table of just the unfiltered rows 3. Use the Summarize function on that table to get my list of displayed axis values ldf_str = ldf << Get Where Clause; //ldf = reference to local data filter in gb window clause = Parse(Regex(ldf_str, "Select Where$(.+)$$", "\1")); r = dt << Get Rows Where(Eval(clause)); //dt = ref to main datable dtsub = dt << Subset(//Invisible, Selected Rows(0), Rows(r), Selected columns only(0) ); Summarize(dtsub, xlev_lis = By(:age, :sex)); This method seems unwieldy though. Is there a command that can "get" the axis values from a graphbuilder reference or an axis box?

nikles · ‎06-16-2020

Thanks. That was exactly what I needed.

nikles · ‎06-15-2020

Thanks nascif_jmp. I'm actually still on JMP14, so I'm not yet familiar with Hover Label Extensions. I can switch to 15, but old habits... Quick question though...I'm looking for some method to programmatically read the variables from a window containing a graphbuilder plot. However, from its name it sounds like the Hover tool would require the user to interact with the window by hovering over various items. Is there a way to have this done automatically by a script? And can it be done if all I have is a reference to the window in which the gb plot exists (not a reference to the gb itself)?

nikles · ‎06-15-2020

Hi Jim. Thanks for your reply. Unfortunately the problem I have is that my script will not have any knowledge of the "gb" ref in your example. The only thing my script would be able to reference is the window and the subsequent display boxes. If I could somehow get the reference to gb however, then yes, your method is precisely the way I'd prefer to go.

nikles · ‎06-14-2020

Hi. Is there a method to get the variable names & roles used in a graph builder window? Suppose I have a window containing a graph builder plot. Is there a way to programmatically determine which variable was used for X, Y, Xgroup, Ygroup, etc? For example, suppose the user created a plot in gb. Now they wish to run some generalized script on the plot (e.g. a graphics script). The script needs to know what variables were used in which roles without the user telling it. Suppose I start with this plot: Names Default To Here(1); dt = Open("$SAMPLE_DATA/Big Class.jmp"); gb = Graph Builder( Variables(X(:height), Y(:weight)), Elements(Points(X, Y), Smoother(X, Y)) ); Now I wish to run a script on the resulting plot. My script is unaware of what is contained in the plot. Here are 2 things I tried but did not work: Names Default to Here(1); //Try to get script used to create plot win = Window(1); //gets most recent window clicked on gb = win << Report; //DNW: returns missing (win is not a platform) scr = gb << Get Script; //from here I could parse the string to determine the roles //Another stab: try to read the XML win = Window(1); str = win << Get XML; //This gives me the XML of the window, but does not provide the actual variable names used in the roles. If I could figure out how to convert the window reference into a report, I think I could get that first method to work. Any ideas?

nikles · ‎11-20-2019

Thanks Tex. I wasn't aware you could get a script from a report. That opens several doors. Although I still have hope for there to be some method to do this via XPath.

nikles · ‎11-19-2019

Hi. My question is if there is a simple way to navigate a display tree to get to the (x) AxisBox of a plot, if it's unknown whether other (y) AxisBoxes precede it in the tree. I'm considering possibly using Xpath or some other method. In my script, I create a window for my results, including a distribution. User has option of including histogram, outlier box plot, and/or normal quantile plot. Example: Names Default to Here(1); dt = Open("$SAMPLE_DATA/Big Class.jmp"); //User controllable options to plot histogram, outlier box plot, or normal quantile plot histopt = 1; outlopt = 1; nqplopt = 1; //Make results window Try(Window("My Window") << Close Window); win = New Window("My Window", TextBox("Important text info"), LUB = LineUpBox(NCol(1), Spacing(10,10)) ); //Add distribution to it. Eval(EvalExpr( LUB << Append( dist = Distribution( Column("Height"), Histogram(Expr(histopt)), Outlier Box Plot(Expr(outlopt)), Normal Quantile Plot(Expr(nqplopt)), Capability Analysis(0), Quantiles(0), Vertical(0), Stack(0), Automatic Recalc(0) ); ); )); win << Show Tree Structure; With all 3 plot options selected as above, the resulting tree structure shows me that the (y) AxisBox for the Normal Quantile Plot is titles "AxisBox(1), and the (x) AxisBox is titled "AxisBox(2)". If I turn off the normal quantile option and run the script again, the (x) AxisBox is now titled "AxisBox(1)". So unfortunately I cant refer to the Axis Box simply by its name, plus I don't like doing that as display box names can change. I do not know XPath well enough to tell it how to look for only x-axis Axis Box, but I'm wondering if there is some method like: win << XPath("\\AxisBox() = "Xaxis"); Or something akin to that?

nikles · ‎11-14-2019

Thanks Craige. Your first assumption was correct: I wish to remove the literal values of 2 and 5 from the vector containing 1,2,3,4,5. Another example to help clarify, I wish to remove this subset vector from largevect: largevect = [12, 189, 3, 78, -2] subset = [189, -2] dsired result = [12, 3, 78] I've not used AA's much, but I like this as a solution. Thanks for the help!

nikles · ‎11-10-2019

Hi. I have a large vector (i.e. 1-D matrix) of values, from which I'd like to remove a subset of the values, where the subset contains a non-sequential set of values from the original vector. For example: largevect = [1,2,3,4,5] subset = [2,5] desired result = [1,3,4] I'm wondering if a simple command or set of commands exists that could do this in JSL? I could do this with a For Loop, but wondering if there's a more elegant or faster method?

nikles · ‎11-04-2019

Thanks. This works, but for the life of me, I could not find this in the Scripting Guide.

nikles · ‎11-01-2019

Hi. I'd like to know if there's a way to remove a "Compare Distributions" from an existing Distribution platform output using JSL. For example: Names Default To Here(1); dt = Open("$SAMPLE_DATA/Big Class.jmp"); obj = Distribution( Column(:Height, :Weight), Fit Distribution("All") ); //These are things I've tried to remove the CmpDist for the Height Distribution, but did not work: obj[1] << Remove Fit; obj[1] << (Fit Handle[1] << Remove Fit); obj[1] << Fit Distribution(0); I want to completely remove the "Compare Distributions" outline box and its contents, along with the fit curve in the plot. IOW, as if I'd never issued the Fit Distribution("All") command at all for that distribution. IOOW, as if I went to the red down arrow for the Height distribution > Continuous Fit and unchecked "All". I have not yet attempted to delete the display boxes from the report window via brute force. Hoping there's a simpler option?

nikles · ‎10-31-2019

Thanks! I'll give that a try! Plus - was not aware about the Ignore Platform Prefs command. That makes life a lot easier.

Date Registered	‎03-04-2013 04:11 PM
Date Last Visited	yesterday
Total Messages Posted	82
Total Kudos Received	30

Online Status	Offline
Date Last Visited	yesterday