Hi there
I want to exclude some values from d when it is bigger than the width .... I want the program to just exclude the values that do not fulfil the condition, I used this program ,but it shows strange error ...I'm sure both of them have the same units
How to fix this ?
ACCEPTED SOLUTION
Accepted Solutions
You may also use a generic utility function to do the job:
You may then use either
or in one go
You have to decide what to return in case all elements have to be deleted. As Prime does not provide an empty array you may chose 0 or NaN or, like I did, a 1x1 array containing NaN (so I can still use the transpose operator even in that case).
Here some examples of usage
Delete the word "rows", just use if d[i < width
A simpler way probably would be to just use
or
Caution: This will fail, if ALL values in d are smaller than width or if ALL are larger or equal.
EDIT: See a more generic function which also deals with these special cases in my answer below.
You may also use a generic utility function to do the job:
You may then use either
or in one go
You have to decide what to return in case all elements have to be deleted. As Prime does not provide an empty array you may chose 0 or NaN or, like I did, a 1x1 array containing NaN (so I can still use the transpose operator even in that case).
Here some examples of usage
Neat.
My preference is to return 0, as that way, it is clear that the result is not a vector and can be conveniently tested using a boolean. In addition, there are cases where [NaN] might be a valid (vector) result.
To deal with the transpose problem, I use a nifty transpose function, usually defined as T (in Constant Math Style), which transposes array inputs but leaves scalars alone.
I sometimes also use it (or suitable variants) under vectorization. I have other variants for dealing with nested vectors when the display gets too tall or difficult to interpret.
Oddly, I also find it more keyboard-friendly, as typing "shift-T shift-9" comes to my fingers more readily than positioning the cursor where needed and typing "ctrl-shift-t," probably because it requires less hand movement.
Stuart
@Werner_E wrote:
Nice, I like your many little helper functions which on first sight do not really much but at the end prove to be very handsome and helpful.
Wish one day I could succeed in creating my own library of utilities, being able to find it when I need and remember the function names i had chosen, rather than re-inventing the wheel from anew each time....
I've now reached the stage where my new wheels get similar names. Plus, I have a "functions" Area in my default worksheet, where I keep my most useful or frequent functions.
If I get stuck for a name, I browse the MATLAB, Mathematica, Maple, numpy, or other application Help to find something close to what I want.
Here's an extract showing a more ... complex version of T that I use for vertically compacting vectors or nested arrays. It goes with saying that I need to carefully interpret what it's telling me, as it transposes column vectors but not row vectors (which would rather obviate T's purpose).
Naturally, I tinker with T when it doesn't quite meet my particular needs for viewing a nested array result.
Stuart
After a few minutes of thought, something started bugging me about exclude and select. There was something very familiar about the problem, but I couldn't put my finger on it, just vague sense of deja vu. So, I grabbed a cup of coffee and cast my mind back into deep time, and finally reached the dizzy depths of last week.
I'd been busy upgrading some M11 functions to Mathcad Prime 10, so they harkened to an era before trim. Consequently, I'd had to adopt a different approach, namely using a standard functional programming filter function. filter also solved another problem I'd hit more frequently than not: filtering matrices and ranges, which, as they stand, your exclude and select don't handle. However, it's an easy fix: use my ubiquitous vec function to flatten the matrices and expand the ranges. I've got another variant of filter that accepts a vector of indices and returns the corresponding elements (although I usually used a function called pick to do that job; sometimes, universality can go a little bit too far!).
But that wasn't the end of it, not by a long chalk. I'd become fed up with non-empty empty array stand-ins (a regular occurrence).. So, I reinvented the wheels known as stack and augment, called them stk and aug, and made them pretend they knew what an empty array was. It was quite pleasant not having to worry about initialising arrays before loops.
which then let me rewrite vec and filter accordingly,
And here are a few functions from another library worksheet I've got on the boil.
I haven't forgotten about the Expression Type Functions worksheet. However, answering your questions sent me off down a few other rabbit holes, and then I had some other ideas that needed addressing before I forgot them, and, etc, etc ... Fairly rapid progress is being made, though. and it might even be ready for publication in a couple of days.
Stuart
Nope, nothing to do with Excel. That all works the same.
I mean...
versus
If you wanted an easy way to turn a range variable into a vector, you can borrow this suggested small program from CS416474.
@DJNewman wrote:
If you wanted an easy way to turn a range variable into a vector, you can borrow this suggested small program from CS416474.
The function Range2Vec is similar to many others that have appeared over the years, including my vec family of flatteners.
I have a problem with the name "Range2Vec".
Firstly, it takes more typing and use of the shift key. Think of the RSI ...
Secondly, functions that use this for-loop iteration over a variable do more than just convert ranges. As I said in a previous message, they'll convert matrices into vectors as well. My recursive version will flatten arbitrarily nested arrays into vectors.
The reason I chose the name vec is because a) it's short, easy to type, and easy to remember; b) it conforms with the mathematical operation of the same name, and as Mathcad is a mathy app, I feel that mathy names ought to apply where possible.
https://en.wikipedia.org/wiki/Vectorization_(mathematics)
In mathematics, especially in linear algebra and matrix theory, the vectorization of a matrix is a linear transformation which converts the matrix into a vector. Specifically, the vectorization of a m × n matrix A, denoted vec(A), is the mn × 1 column vector obtained by stacking the columns of the matrix A on top of one another:
I think it was Tom Gutman, one of the Mathsoft era Mathcadders extraordinaire, who first proposed the basic method, using the name ravel, possibly from some of the mathematically-based programming languages, such as APL,
Ravel (,) is a primitive function introduced in APL\360 which returns the ravel of an array. In the APL array model, an array's ravel is the vector containing all its elements in ravel order. It is equivalent to reshaping an array using its bound for the new shape. Reshaping the ravel using the original array's shape restores that array.
In some APLs an axis may be specified for Ravel in order to combine only some axes of an array, or insert a length-1 axis.
The name "ravel" references the process of undoing woven or knitted fabric, thus removing its structure and rendering it linear.
I forgot that people use index bases other than zero, so here are ORIGIN-independent versions of flat and recursive vec ...
Have I mentioned that Mathcad could really do with an empty array value? 😉
Stuart
From Wikipedia, the free encyclopedia Vectorization may refer to: Array programming, a style of computer programming where operations are applied to whole arrays instead of individual elements Automatic vectorization, a compiler optimization that transforms loops to vector operations Image ...
@YA_10963798 wrote:
What is the undocumented trick ?Do you mean calling data from excel ?
No, it was to do with how some versions of Mathcad Prime dealt with the result of evaluating (= operator) a range variable when you defined it.
In MP10, k is still a range variable, but in MP9 k would be a vector and thus transposable. I rarely used the trick and, although I wasn't looking for it, I didn't notice any examples in your worksheets. It's an easy enough fix, though; if you need to use the name "k", just define the range variable as, say, "kk" and iterate over it to produce k as you would any other vector/array. Alternatively, use a simple version of vec to do the job without any name changes.
Of course, I could be wildly wrong about the undocumented trick - there are so many of them! 😉
Stuart
Hi Werner @Werner_E
I'm trying to show the results the same way you did , but error occurs . It didn't work once I add "row" , though I used your way and Stuart way ,
this is result3
I need to show the rows of the results and then draw it the same way you draw the this one , but only with the selected values
can you show me how ?
This is now not a question concerning "Exclude some values", isn't it.
It rather a question belonging to the thread about the nearest point.
I was not ware that you also needed the rows number of that nearest point because you did not specify it in your intial posting there. You just said that you need the nearest point and the distance.
Its not difficult to modify the functions to also provide the index of the point.
Here done with the first 'table' (function 'getNearest') I provided in that thread
Werner, I'm so sorry to ask so many question...I promise this is my last question in this thread
I'm trying to write a new program to show the selected points ... the word "result 3" is the exclued values... but obviously what I wrote is wrong
I'm trying to write the same program , so then I can plot it as we draw it before ..
I want to show the selected values in different color in this plot
The plot I made out of fun was meanly meant to be used for my demo data with just a couple of points in each vector. It should demonstrate that its possible that two or more points in A have the very same minimal distance from a point in C as the question as to what to do in this case was not (and still isn't) answered. My "table" used just the first of the multiple point which may be possible, "table2" would show all of them which is the reason why the "A point list" in table2 are vectors as there could be more than one nearest point for a specific C-point.
So the reason for the plot was not only to show the points but mainly to connect corresponding points with straight lines.
This did not make much sense with the original data so I zoomed in by adjusting the axis limits manually (Prime unfortunately does not offer a zoom facility).
Without zooming it would look like this. The connections between the corresponding points would not be seen.
I also noticed that with the Excel data you provided there is no single point in C with more than only one nearest A-point. If this is always so for sure or if it suffices to use the first one if there would be more than one point with different input data, I guess that "table" is preferable over "table2" because its not nested and so makes for a nicer display. Of course there may be reasons for rather using a structure as in "table2" or as the one Stuart had provided. Its up to your choice.
Concerning the plot I am not sure what you would like to plot.
I have not looked into detail in the sheet you posted last as it was a bit too confusing for me with all the different suggestions for the nearest points mixed up.
I look at result3 now and its just a vector of 1308 distance values. So what exactly would you plot here? I am not sure what you would like to see in the plot.
Guess you would first decide upon the data format for A and C.CPT you want to use (either still nested or flattened), decide on one of the methods to create the nearest point list and then chose a way to extract whatever points or distances you need.
Maybe cleaning up the worksheet in this respect would help.
Form the picture you posted its not clear to me what your function "new" is supposed to create with its three nested for-loops.
