Input argument "w" is undefined
in [Matlab]
|
Prev: A possible bug in Windows Server 2008?
Next: reshape
From: Peter on 24 Jul 2010 09:23 New to Matlab, so a beginner question which after many hours I still can't find the answer to: Defined a function, "planck": ======= function i = planck( w,t ) % planck function for one value of wavelength, w and temp, t % result in spectral intensity i = 3.742E8.*(pi.*w.^5.*exp((1.439E4./(w.*t))-1)).^-1 end ======= and it works fine when I call it from the command line, like planck(0.5,5780); but I do notice - strange and does this mean a problem? - that the output is duplicated (and shouldn't the ";" suppress this anyway??): "i = 7.1276e+007 ans = 7.1276e+007" Anyway, then I defined a script to run and call this function. (Cut down version of the original idea here, because I kept "winding it back" to eliminate the error). Script: == for w=.02:.02:100 y = planck(w,290); end ==== But when I "play" the script the command line says: ===== ??? Input argument "w" is undefined. Error in ==> planck at 4 i = 3.742E8.*(pi.*w.^5.*exp((1.439E4./(w.*t))-1)).^-1 ===== I hope someone can help, this is probably so easy, but I cannot find documentation with enough examples to supply the solution. Thanks
From: dpb on 24 Jul 2010 09:39 Peter wrote: > New to Matlab, so a beginner question which after many hours I still > can't find the answer to: > > Defined a function, "planck": > > ======= > function i = planck( w,t ) > % planck function for one value of wavelength, w and temp, t > % result in spectral intensity > i = 3.742E8.*(pi.*w.^5.*exp((1.439E4./(w.*t))-1)).^-1 > > end > ======= > and it works fine when I call it from the command line, like > planck(0.5,5780); > > but I do notice - strange and does this mean a problem? - that the > output is duplicated (and shouldn't the ";" suppress this anyway??): > > "i = > > 7.1276e+007 > > > ans = > > 7.1276e+007" You don't have a semicolon inside the script; the first "i = " is coming from that. Matlab echoes everything unless you suppress it (which is handing for debugging, just leave the semicolon off a line to see it's effect temporarily). You also do _NOT_ want to write a function named "i" -- i (and j) are defined functions in Matlab for the imaginary unit vectors and great confusion arises from redefining them. > Anyway, then I defined a script to run and call this function. (Cut down > version of the original idea here, because I kept "winding it back" to > eliminate the error). > > Script: > == > for w=.02:.02:100 > y = planck(w,290); > end > ==== > > But when I "play" the script the command line says: > > ===== > ??? Input argument "w" is undefined. > > Error in ==> planck at 4 > i = 3.742E8.*(pi.*w.^5.*exp((1.439E4./(w.*t))-1)).^-1 > ===== .... Hmmm....I don't see this otomh... Think need to fix the previous comment re "i", add the semicolon in the function definition, then do a "clear all" and try again. I'm thinking there's a problem w/ caching a previous version or somesuch so that the function being evaluated isn't actually the latest version you think is being called or a similar problem. If that doesn't cure it; then repost w/ latest version and script in context. BTW, you recognize your script as given will overwrite the "y" each time through the loop I presume? --
From: Peter on 24 Jul 2010 09:59 dpb <none(a)non.net> wrote in message <i2eqk5$1u4$1(a)news.eternal-september.org>... > Peter wrote: > > New to Matlab, so a beginner question which after many hours I still > > can't find the answer to: > > > > Defined a function, "planck": > > > > ======= > > function i = planck( w,t ) > > % planck function for one value of wavelength, w and temp, t > > % result in spectral intensity > > i = 3.742E8.*(pi.*w.^5.*exp((1.439E4./(w.*t))-1)).^-1 > > > > end > > ======= > > and it works fine when I call it from the command line, like > > planck(0.5,5780); > > > > but I do notice - strange and does this mean a problem? - that the > > output is duplicated (and shouldn't the ";" suppress this anyway??): > > > > "i = > > > > 7.1276e+007 > > > > > > ans = > > > > 7.1276e+007" > > You don't have a semicolon inside the script; the first "i = " is coming > from that. Matlab echoes everything unless you suppress it (which is > handing for debugging, just leave the semicolon off a line to see it's > effect temporarily). > > You also do _NOT_ want to write a function named "i" -- i (and j) are > defined functions in Matlab for the imaginary unit vectors and great > confusion arises from redefining them. > Ok - thanks I see your point and changed both of these - new function: ===== function ri = planck( w,t ) % planck function for one value of wavelength, w and temp, t % result in spectral intensity ri = 3.742E8.*(pi.*w.^5.*exp((1.439E4./(w.*t))-1)).^-1; end ======= Once again it works fine, but now no echo. > > > > Anyway, then I defined a script to run and call this function. (Cut down > > version of the original idea here, because I kept "winding it back" to > > eliminate the error). > > > > Script: > > == > > for w=.02:.02:100 > > y = planck(w,290); > > end > > ==== > > > > But when I "play" the script the command line says: > > > > ===== > > ??? Input argument "w" is undefined. > > > > Error in ==> planck at 4 > > i = 3.742E8.*(pi.*w.^5.*exp((1.439E4./(w.*t))-1)).^-1 > > ===== > ... > > Hmmm....I don't see this otomh... > > Think need to fix the previous comment re "i", add the semicolon in the > function definition, then do a > > "clear all" > > and try again. I'm thinking there's a problem w/ caching a previous > version or somesuch so that the function being evaluated isn't actually > the latest version you think is being called or a similar problem. > > If that doesn't cure it; then repost w/ latest version and script in > context. > > BTW, you recognize your script as given will overwrite the "y" each time > through the loop I presume? > > -- I cleared all and ran my script (no change): == for w=.02:.02:100 y = planck(w,290); end ==== But same error as before: ==== ??? Input argument "w" is undefined. Error in ==> planck at 4 ri = 3.742E8.*(pi.*w.^5.*exp((1.439E4./(w.*t))-1)).^-1; ===== And what I was trying to do was create an array, y, the same size as w - not replace y with a new value each time through.
From: dpb on 24 Jul 2010 12:54 Peter wrote: .... > I cleared all and ran my script (no change): > == > for w=.02:.02:100 > y = planck(w,290); > end > ==== > > But same error as before: > > ==== > ??? Input argument "w" is undefined. > > Error in ==> planck at 4 > ri = 3.742E8.*(pi.*w.^5.*exp((1.439E4./(w.*t))-1)).^-1; > ===== I built a function function ri = planck( w,t ) % planck function for one value of wavelength, w and temp, t % result in spectral intensity ri = 3.742E8.*(pi.*w.^5.*exp((1.439E4./(w.*t))-1)).^-1; and then, >> for w=.02:.02:100,y = planck(w,290);end >> y y = 0.0197 >> Works here. One thing I note is an unmatched "end" statement in your function listing which makes me think there's something not being shown here. As written otherwise should work afaict. > And what I was trying to do was create an array, y, the same size as w - > not replace y with a new value each time through. As you wrote it, you don't need the loop for the one variable--that's what the .*, ./ operators do is to evaluate on an element-by-element basis to eliminate loops; Matlab's strength. Now if you want to do both variables simultaneously, you'll need to build a grid or somesuch, but as you've written it, planck() will return the vector directly. >> y = planck([.2:.2:100], 290); >> length(y) ans = 500 >> If you really do want to call it in a loop on a value-by-value basis then that would look something like y = zeros(length([.2:.2:100])); % preallocate idx = 1; for w=[.2:.2:100] y(idx) = planck(w,290); idx = idx+1; end --
From: Jan Simon on 24 Jul 2010 13:06
Dear Peter, the description of your Porblem sounds correct. The source code you've posted looks correct. The error message tells you, that it is not correct. Scientific conclusion: You run other code than you have posted. Ideas for a solution: Do you have more than one function called planck? which planck.m Have you saved the file planck.m after editing? Let the debugger show you, what's going on. Set a breakpoint in your function "planck" and run it again line by line. Good luck, Jan |