About nathan-clark

nathan-clark

I would like the ability to update the delete columns function so that it can "delete formula" or "remove ref" automatically when 'errors' occur. Currently I can manually delete a column and JMP will notify me if that column is referenced by a formula in another column, giving me the options of how to proceed. But in script, if I code the deletion of a column referenced by another, the column won't delete and I just get a note in the log. In a dynamic setting I may not know which, if any, formulas exist and if they reference columns which may be deleted. So I would like a way to automate that and either do nothing, remove references or remove formulas from the JSL interface. Thanks!

nathan-clark

This is kind of old, but I have an addition to this and I figure I'll try here first 🙂 What if I don't know the columns with formula that will be referencing the column i want to delete. I am trying to make something dynamic, and I would like to precode it so it wilkl automatically delete the formula if there is a conflict (just like if i did it manually), but I can't find any syntax that allows for that. the "easiest" method is to just remove all formulas so it's never an issue, but that seems a bit more extreme than it needs to be.

nathan-clark · ‎04-03-2020

I would like to suggest the ability to force MFI to combine files together, even if it finds differences. Use Case: I am combining data from two instrument platforms where the data is set up identically, but the differences are enough that MFI separates the data into two tables if I have both instrument platforms. I am aware of all pertinent differences, and just want to leverage MFI for high quality, quick combination. I would like a radio button to 'force into one file' along with "Stack Similar Files" and "One Table for Each File". While JMP does an amazing job most of the time, sometimes I know better than JMP 😉

nathan-clark · ‎04-03-2020

I would like to suggest the ability to simply 'pick files' from the list in a folder rather than figuring out a filter to apply or moving some files into separate folder. Use Case: I have a dozen random-ish files in a folder and I want to combine 4 of them. Rather than finding a filter which would allow only those 4 files to be shown, I want to be able to ctrl+click them to build the list manually. (as i click the name populates in the 'select by name or extension box?)

nathan-clark · ‎03-22-2020

I have seen that ... while it's a good generalization of all the JMP Pro features, I think it's more of an impact to be able to see the pro feature when you want / need it most. For instance, JMP offers some levels of simulation in Standard, and more options with platforms in Pro. Rather than remember all the instances where this may occur and when I may want it, I'd rather be able to be in the middle of analysis and see what I can and can't do directly with this data set in front of me. The more I am "burned" by seeing additional options I can't use because I don't have Pro yet, I can much more easily make the case why I need pro and list specific metrics toward how often I would be using it both for the times I was "burned" by not having it as well as all the grayed out features I may see in passing that would be useful with any given analysis but can't use.

nathan-clark · ‎03-20-2020

I would like JMP standard to have all of the options JMP pro has ... however, those options are grayed out and have a little 'pro' icon or something. It would make it SO much easier to discern between colleagues, webinars, presenations, etc what I can do with standard or what I can't. Moreover, it's a very obvious way to show people who use Standard what they are missing in JMP Pro!

nathan-clark · ‎03-20-2020

I would love to see this as a default also. I have found a couple hacks that give me this on my main JMP as well as any JMP I use a shortcut with, but I'd rather JMP do the work than me 🙂

nathan-clark · ‎02-11-2020

I would like propose an update so that a preference (or some other method) is in place for how outline boxes, especially the default one, are handled during copy / paste operations - for graph builder, specifically (although it could be merged into other platforms). This preference would allow the user to control whether or not the default outline box was copied / pasted when the user selects it for copying. In JMP 15, an feature was implemented which automatically removes the outline box from the image when copying / pasting from Graph Builder. This feature removes the need to manually select each graph with the control key to copy / paste when the titles are effectively part of the graph builder output. Additionally, JMP implemented it so that editing the outline box would force keep it when copying / pasting so user adjustments were maintained. However, if a user wants to keep the outline box(es) it is just as annoying to change each title so it is maintained as it is to manually control click each graph to avoid copying the outline box. A preference would allow users to control their destiny and be the best of both worlds.

nathan-clark · ‎12-10-2019

Thanks! I caught up with the conversation and I think you guys are on the right track. I look forward to seeing what is merged in!

nathan-clark · ‎12-10-2019

@Justin_Chilton, I was hoping you (or someone else) may be able to give me a hand with jsl hamcrest question. I am looking to extend the reporting of the tester, but I am not quite sure how to navigate the classes to do so. I found the UtCollectiongReporter and was able to have a blank report end up at the end of my test by adding this line at the end: UtCollectingReporter:show report(); However, none of the successes/failres/etc were included. How to I adjust the tool so this reporting tool is filled with information? Thanks!

nathan-clark · ‎12-10-2019

I was just thinking about this the other day, @vince_faller. Have you found an answer yet? I was looking into using the Compare Data Tables platform, but didn't see an output I could use in a unit test.

nathan-clark · ‎04-01-2019

Thanks for the idea @Craige_Hales! I had tried the <<write "bye" before, but without the <<writeEoF ... so perhaps that would make a difference. I had thought of other FTP clients, but this script would be used across a larger user base on which I can't control the FTP client piece. If I find a solid thing that works, I'll post it here, but I'll def try some of your ideas!

nathan-clark · ‎04-01-2019

Awesome you got a solution to work @DS ! That bit of code has helped me out numerous times. And @ian_jmp, that's som really good stuff with SubstituteInto(). I can definitely see how it could be more dynamic and controlled with that. I'll need to add it to my tool box 🙂

nathan-clark · ‎03-29-2019

I am in a situation where I use Run Program() to send files to an FTP/MTP site. By and large everything works great, but sometimes there's a kink in the connection and the program lags indefinitely. I have recreated this problem from the command line and it's basically a lost connection type event. But I need to Ctrl-break out of CMD to stop it. In JMP I want to put in a timeout to reduce the risks, however I haven't been able to get one to work well. Currently I build the timeout into the While(!(x << isReadEof), Wait(0.1)) command such that x is set to 0, breaking the Run Program. But when I do this JMP will often freeze for a short or longer time to fix (or while something else is fixed) background stuff that had been going on. Is there a way to force close Run Program() as well as the process it invoked w/o risk to freezing JMP? Thanks!

nathan-clark · ‎03-29-2019

The best way I have found to do this is using Eval() and EvalExpr() ... basically, the variables in the script are expressions and eval expression evaluates them so the final evaluation of your table script has the actual values, not the variables. Below is a quick example I grabbed where I created a table script and it will insert the specific column names as opposed to the variable (parseUnique). parseUnique is created by inserting a text string for the column name into the :Name() function so column names work better. I don't need to worry about variables I create within the expression (tblName) but any variables that come from outside sources become expressions with in the function. parseUnique = EvalInsert(":Name(\!"^uniqueCol^\!")"); Eval( EvalExpr( dt_algorithm << New Script("Make justResults", tblName = currentdatatable()<<Getname; Try(newName = Munger(tblName,1,"_progCurves",""),newName = tblName); datatable(tblName) << SelectWhere(Expr(Parse(parseUnique)) != Lag(Expr(Parse(parseUnique)),1)); datatable(tblName) << Subset( output table name( newName|| "_justResults" ) ); datatable(tblName) << select all rows; datatable(tblName) << invert row selection; ) ) );

nathan-clark · ‎11-26-2018

Thanks for all the answers so far! The OLE solution may be the best bet, although it's not very ideal ... but then again, I didn't expect a slam dunk answer 🙂

nathan-clark · ‎11-26-2018

Has anyone developed/found a way to convert/share/"compile" JSL into something usable by other languages? Here is my problem. As the developer of the initial code in R&D, I want a way to be able to give researchers code/tools that meet their needs and work with there existing analysis structure (JMP). However, some of these tools get merged into larger suites (for operations/QC, etc) where an existing infrastructure is already in place, not based on JMP. How can I share what my JSL tool does without having to recreate all of the code in a new language to be fully validated again. In my head there would be a button “Convert JSL to Python” (or C++, etc) which would do the work. I doubt it’s that simple, but I was hoping there were tools in existence that would help me turn my JSL into something more digestible for other languages. In many cases the final users aren't JMP users so having JMP on their machines isn't a given (however, if that's the only way to make it work, I want to explore all options). Thanks!

nathan-clark · ‎04-06-2018

@Justin_Chilton Thanks! I'll give that a try.

nathan-clark · ‎04-04-2018

@Justin_Chilton Thanks again for this addin. I am finally getting up to speed with the new version and taking advantage of some of your improvments on what I was doing manually 🙂 My one ask, and I'm not sure if this is possible, is to be able to build several addins at once. I know the bottom panels for information and directories are addin specific, but perhaps those could disappear if more than one addin is selected. Then all selected could be DEVd/Deployed/etc. I manange over 40 addins right now and I often make updates in batches so being able to do a bunch of all of the addins at once would be amazing!

nathan-clark · ‎11-21-2017

Wow, @Jeff_Perkinson and @Phil_Brown, I came looking for some general stuff on indexing through associative arrays and I got so much more! As a recent convert to associative arrays (so many nested lists in my past 😕 ) I'm trying to use them more often and there must be at least 4 separate ideas in this thread alone I can merge into various sets of my code! Thanks for posting this 🙂

nathan-clark · ‎11-07-2017

That does look like a cleaner solution. Too bad JMP doesn't have this as a canned rounding thing. To be honest, I had no idea there were other kinds of rounding than ... standard? rounding 🙂 Thanks!

nathan-clark · ‎11-07-2017

Ok, that looked horrible.... let's try this for the two formula options: bankerRound = Function({inputValue}, {inputValue,fDiff,cDiff,modX,bankRound}, fDiff = inputValue - floor(inputValue); cDiff = ceiling(inputValue) - inputValue; modX = Modulo(round(inputValue,0),2); if(cDiff == fDiff & modX == 0, bankRound = Ceiling(inputValue) , cDiff == fDiff & modX != 0, bankRound = Floor(inputValue); , bankRound = Round(inputValue,0); ); Return(bankRound) ); inputValue = Empty(); // place column name here fDiff = inputValue - floor(inputValue); cDiff = ceiling(inputValue) - inputValue; modX = Modulo(round(inputValue,0),2); if(cDiff == fDiff & modX == 0, bankRound = Ceiling(inputValue) , cDiff == fDiff & modX != 0, bankRound = Floor(inputValue); , bankRound = Round(inputValue,0); );

nathan-clark · ‎11-07-2017

I was recently asked about banker's rounding and didn't find anything in the JMP documentation. I wrote the quick function below (as well as a formula option), but was curious if 1) JMP has a canned method for doing banker's rounding and 2) what other/better methods may be useful vs. mine. http://wiki.c2.com/?BankersRounding bankerRound = Function( {inputValue}, {inputValue, fDiff, cDiff, modX, bankRound}, fDiff = inputValue - Floor( inputValue ); cDiff = Ceiling( inputValue ) - inputValue; modX = Modulo( Round( inputValue, 0 ), 2 ); If( cDiff == fDiff & modX == 0, bankRound = Ceiling( inputValue ), cDiff == fDiff & modX != 0, bankRound = Floor( inputValue ), bankRound = Round( inputValue, 0 ) ); Return( bankRound ); ); inputValue = Empty(); // put column name here fDiff = inputValue - Floor( inputValue ); cDiff = Ceiling( inputValue ) - inputValue; modX = Modulo( Round( inputValue, 0 ), 2 ); If( cDiff == fDiff & modX == 0, bankRound = Ceiling( inputValue ), cDiff == fDiff & modX != 0, bankRound = Floor( inputValue ), bankRound = Round( inputValue, 0 ) );

nathan-clark · ‎02-09-2017

Thanks to everyone's comments... I figured out what the problem was and why Deletesymbols() wasn't working. It turns out that one part of my scripts turned on the Here namespaces (Names default to here(1)) ... being part of an include made the whole batch of scripts in the Here namespace which didn't work with other scoping I had done. Deletesymbols() didn't work because the variables I needed to delete were actually stored in the Here namespace now, and protected. I was able to fix this by forcing the Here namespace to go away at the start of my 1st script so when it was brought back by the 2nd script all the failures would occur on demand. In the very end, the two scripts came from different sources and a lot of scoping work needed to be done to get everything working again. But I now know more about how the Here namespace can be a blessing and a curse 😄

nathan-clark · ‎02-08-2017

Have you tried to build the addin yourself outside of the addin builder? Is there an error in the log when you try to run the addin from the addin menu?

nathan-clark · ‎01-30-2017

Sadly, DeleteSymbols() doesn't work. Much like your response I believe it should have worked. I may need to contact JMP support. Perhaps DeleteSymobols() isn't doing what it should be doing?

nathan-clark · ‎01-29-2017

There are a bunch of them (variable, namespace assignment, etc) and not worth bringing up here. All I'm looking for is a script that will completely clean all variables, namespaces, globals, cache, etc. so the state is as clean as an intial start. I know the reason(s) the script isn't working ... i just dont want to completely restart JMP every time I want to test.

nathan-clark · ‎01-29-2017

I have a script that I'm troubleshooting and the issue(s) deal with scoping. The problem is that I have to completely shut down JMP and restart it to test my changes ( deletesymbols() and deletenamespaces() don't clean things up enough). 1) Start JMP 2) Run script, get error 3) Close data tables 4) Run script again, script works fine. 5) Close JMP then restart JMP 6) Run script, get error Clearly something is declared, in place, assigned, etc that allows the script to work on the 2nd time. I'd like some script to pretty much clear whatever JMP has going on so it's about as fresh as possible (just like a full restart of the program). Thanks!

nathan-clark · ‎01-24-2017

@vishwasanj, have you tried building the addin manually? You should be able to just add the files you need into a .zip file and then rename the zip file .jmpaddin. It should give you the same installation. If that works and the addin builder does not it could be that where you are pointing the addin builder to make the addin from isn't quite correct. You can also open an addin file with a zip reader to ensure that all the files are correct in the .jmpaddin file.

nathan-clark · ‎01-06-2017

I LOVE this addin! I use it all the time. Two updates that would be awesome would be changing the default table location to be outside of a specific JMP version (I routinely operate in two versions of JMP) and being able to build more than one addin at once. I was actually looking into learning to build something like this via python when I was shown this, so I'm super excited about it! - Thanks 🙂

Date Registered	‎09-27-2016 02:30 PM
Date Last Visited	2 weeks ago
Total Messages Posted	40
Total Kudos Received	28

Online Status	Offline
Date Last Visited	2 weeks ago