Foreword

This document contains a list of some Matlab basis functions and compare them with the Scilab equivalent functions if there are. It gives the main differences and gives examples of particular cases. However not all Matlab and Scilab functions are described here. Equivalents for operators and variables can be found at the end of this document. Caution: This document is not dedicated to explain how functions work. Users can used online manuals of Matlab and Scilab for it. How to use this HTML guide: To have more details about functions and examples, clic on the "triangles" just at the left of Matlab function name. To have a more detailed index, clic on the letters and keywords.

	Matlab	A

abs Absolute value and complex magnitude abs

acosh Inverse hyperbolic cosine acosh

acos Inverse cosine acos

acoth(A) Inverse hyperbolic cotangent atanh(1 ./A)

In Matlab y=acoth(x) and Scilab y=atanh(1 ./x), for real elements of x outside the domain [-1,1], the complex part of Scilab y value is the opposite of Matlab y value. See atanh/.

acot(A) Inverse cotangent atan(1 ./A)

acsch(A) Inverse hyperbolic cosecant asinh(1 ./A)

acsc(A) Inverse cosecant asin(1 ./A)

all Test to determine if all elements are nonzero and

Matlab all function can work with complexes, what Scilab and can not, so a call to abs function can be necessary when translating from Matlab to Scilab. ▹ B=all(A) ↔ B=and(A): If A is a matrix, all(A) is equivalent to all(A,1) in Matlab whereas in Scilab and(A) is a logical AND of all elements of A. If A is a multidimensional array then Matlab treats the values along the first non-singleton dimension, but Scilab returns logical AND of all elements of A. ▹ B=all(A,dim) ↔ B=and(A,dim): In Scilab dim=1 is equivalent to dim=''r'' and dim=2 is equivalent dim=''c''. In Matlab, dim can be greater then the number of dimension of A (in this case, B=A), in Scilab you will get an error message. y = all([1,1,0;1,0,1]) y = [1,0,0] y = all([1,1,0;1,0,1],1) y = [1,0,0] y = and([1,1,0;1,0,1]) y = %F y = and([1,1,0;1,0,1],1) y = [%T,%F,%F]

angle(A) Phase angle atan(imag(A),real(A))

any Test to determine if any nonzeros elements or

Matlab any function can work with complexes, what Scilab or can not, so a call to abs function can be necessary when translating from Matlab to Scilab. ▹ B=any(A) ↔ B=or(A): If A is a matrix, any(A) is equivalent to any(A,1) in Matlab whereas in Scilab or(A) is a logical OR of all elements of A. If A is a multidimensional array then Matlab treats the values along the first non-singleton dimension, but Scilab returns logical OR of all elements of A. ▹ B=any(A,dim) ↔ B=or(A,dim): In Scilab dim=1 is equivalent to dim=''r'' and dim=2 is equivalent dim=''c''. In Matlab, dim can be greater then the number of dimension of A (in this case, B=A), in Scilab you will get an error message. y = any([1,1,0;1,0,1]) y = [1,1,1] y = any([1,1,0;1,0,1],1) y = [1,1,1] y = or([1,1,0;1,0,1]) y = %T y = or([1,1,0;1,0,1],1) y = [%T,%T,%T]

asech(A) Inverse hyperbolic secant acosh(1 ./A)

asec(A) Inverse secant acos(1 ./A)

asinh Inverse hyperbolic sine asinh

asin Inverse sine asin

In y=asin(x), for real elements of x outside the domain [-1,1], the complex part of Scilab y value is the opposite of Matlab y value. y = asin(2) y = 1.5708 - 1.3170i y = asin(2) y = 1.5708 + 1.3170i

atan2 Four-quadrant inverse tangent atan2

Matlab atan2 function can work with complexes (in this case, complex part is ignored), what Scilab atan can not.

atanh Inverse hyperbolic tangent atanh

In y=atanh(x), for real elements of x outside the domain [-1,1], the complex part of Scilab y value is the opposite of Matlab y value. y = atanh(2) y = 0.5493 + 1.5708i y = atanh(2) y = 0.5493061 - 1.5707963i

atan Two-quadrant inverse tangent atan

	Matlab	B

balance Diagonal scaling to improve eigenvalue accuracy balanc

There is no equivalent for B=balance(A) in Scilab, balanc function does not work with only one output value. When used with two outputs, these functions return value in inverse order. [T,Ab] = balance(A) [Ab,T] = balanc(A)

barh Bar histogram horizontal barh

bar Bar histogram bar

beep Produce a beep sound beep

Scilab beep always returns a value but not Matlab function.

besseli Modified Bessel functions of the first kind besseli

Scilab besseli function can work with only one output argument, but the Matlab function can work with two outputs arguments. y = besseli(alpha,x) y = besseli(alpha,x,1) [y,ierr] = besseli(alpha,...) y = besseli(alpha,x) y = besseli(alpha,x,ice),ice = 1 or ice = 2

besselj Bessel functions of the first kind besselj

Scilab besselj function can work with only one output argument, but the Matlab function can work with two outputs arguments. y = besselj(alpha,x) y = besselj(alpha,x,1) [y,ierr] = besselj(alpha,...) y = besselj(alpha,x) y = besselj(alpha,x,ice),ice = 1 or ice = 2

besselk Modified Bessel functions of the second kind besselk

Scilab besselk function can work with only one output argument, but the Matlab function can work with two outputs arguments. y = besselk(alpha,x) y = besselk(alpha,x,1) [y,ierr] = besselk(alpha,...) y = besselk(alpha,x) y = besselk(alpha,x,ice),ice = 1 or ice = 2

bessely Bessel functions of the second kind bessely

Scilab bessely function can work with only one output argument, but the Matlab function can work with two outputs arguments. y = bessely(alpha,x) y = bessely(alpha,x,1) [y,ierr] = bessely(alpha,...) y = bessely(alpha,x) y = bessely(alpha,x,ice),ice = 1 or ice = 2

beta Beta function beta

Matlab beta function can work with only one scalar input an done not-scalar input parameter, but in Scilab both parameters must have the same size. A = 1; B = [1 2 3]; Y = beta(A,B); A = 1; B = [1 2 3]; // So that A and B have the same size tmp = A;A = B;A(:) = tmp; Y = beta(A,B);

bin2dec Returns the integer corresponding to a Given binary representation bin2dec

bitand The AND of two integers bitand

bitcmp The binary complementary of an integer bitcmp

bitget Gets the bit of an integer whose the positon is given in the input argument bitget

bitor The OR of two integers bitor

bitxor Returns the exclusive OR of two integers bitxor

blanks A string of blanks No equivalent

There is no Scilab equivalent function for Matlab box but it can be easyly replaced by Scilab equivalent instructions. A = ['xxx' blanks(20) 'yyy']; A = "xxx"+part(" ",ones(1,20))+"yyy";

box Display axes border No equivalent

There is no Scilab equivalent function for Matlab box but it can be easyly replaced by Scilab equivalent instructions. box on box off box box(h,'on') box(h,'off') box(h) a = gca(); a.box = "on"; a.box = "off"; if a.box=="on" then a.box="off";else a.box="on";end; h.box = "on"; h.box = "off"; if h.box=="on" then h.box="off";else h.box="on";end;

break Terminate execution of a for loop or while loop break

	Matlab	C

case Case switch case

In Matlab expression evaluated can be a cell, in this particular use, all values of cell are considered individually (similarly to a OR). In Scilab it can not be a cell (Matlab particularity can be replaced by others "case" or all switch/case statement can be replaced by a if/then/else statement.).

cat Arrays concatenation cat

cd Change/get working directory cd

Note that cd .. does not work in Scilab, but it does in Matlab. In Scilab you can use cd("..").

ceil Round up ceil

cell2mat Convert a cell array into a matrix cell2mat

cellstr Convert strings vector (or strings matrix) into a cell of strings cellstr

cell Create cell array cell

Note that in Matlab, input can contain complex values (in these cases, only real part of it is taken in account), what Scilab function do not tolerate.

chol Cholesky factorization chol

Scilab chol function can only have one output whereas Matlab one can have two ouputs.

cla Clear current axes No equivalent

▹ cla: Scilab equivalent could be a = gca();delete(a.children); but in this case, all current axes children will be deleted because there is no HandleVisibility property in Scilab graphics. ▹ cla reset: Scilab equivalent is a = gca();delete(a.children);.

clc Clear Command Window clc([nblines])

Note that Scilab function allows to clear only a set of lines above the cursor using clc(nblines). Note that Scilab function can not be used in no window mode under Unix/Linux while Matlab one clears the terminal display as if you were using "clear" command.

clear Remove items from workspace, freeing up system memory clear

Scilab and Matlab clear functions are only equivalent when called using clear or clear name. ▹ clear global ...: Scilab equivalent for Matlab clear global [name] is clearglobal([name]). ▹ clear variables ...: Scilab equivalent for Matlab clear variables is simply clear. ▹ clear keyword ...: For all other keywords, there is no Scilab equivalent for Matlab clear call.

clf Clear current figure window clf

▹ clf: In this case, all current figure children will be deleted because there is no HandleVisibility property in Scilab graphics. ▹ fig_handle = clf: Scilab equivalent is be fig_handle = gcf();clf;. In this case, all current figure children will be deleted because there is no HandleVisibility property in Scilab graphics.

clock Current time as a date vector No equivalent

Even if there no equivalent for Matlab clock in Scilab, it can be emuled as shown in example. c = clock w = getdate(); w(3:5) = []; w(6) = w(6)+w(7)/1000; c = w(1:6);

closereq Default figure close request function delete(gcf())

close Delete specified figure close ↔ xdel ↔ delete

▹ close: If current figure is a uicontrol one, Scilab and Matlab close are equivalent. But if current figure is a graphic window, Scilab equivalent for Matlab close is delete(gcf()). ▹ close(h): If h is a uicontrol figure, Scilab and Matlab close(h) are equivalent. But if h is a graphic window, Scilab equivalent for Matlab close(h) is delete(h). ▹ close('all'): Scilab equivalent for Matlab close('all') is xdel(winsid()). ▹ close(name): There is no Scilab equivalent for Matlab close(name) however, mtlb_close can be an equivalent. ▹ close('all','hidden'): Scilab equivalent for Matlab close('all','hidden') is xdel(winsid()) but Scilab kills all figures even if they are hidden.

colordef Set default property values to display different color schemes No equivalent

▹ [h = ]mtlb_colordef(color_option): Scilab equivalent is fig = gcf();fig.background = -1;[h = fig]; if color_option is equal to "black" or "none" and fig = gcf();fig.background = -1;[h = fig]; else. ▹ [h = ]mtlb_colordef(fig,color_option): Considering fig is a graphics handle, Scilab equivalent is fig.background = -1;[h = fig]; if color_option is equal to "black" or "none" and fig.background = -2;[h = fig]; else. ▹ [h = ]mtlb_colordef('new',color_option): Scilab equivalent is fig=scf(max(winsid())+1);fig.background = -1;[h = fig]; if color_option is equal to "black" or "none" and fig=scf(max(winsid())+1);fig.background = -2;[h = fig]; else.

complex Returns the complex form corresponding to the given real part and imaginary part complex

conj Complex conjugate conj

continue Keyword to pass control to the next iteration of a loop continue

conv Convolution convol

Scilab convol output value is always a row vector while Matlab conv output value is a column vector if at least one input is a column vector. To have a closer result, replace Matlab conv(A) by clean(convol(A)) in Scilab.

cosh Hyperbolic cosine cosh

cos Cosine cos

coth Hyperbolic cotangent coth

cot Cotangent cotg

cputime Elapsed CPU time timer()

csch(A) Hyperbolic cosecant 1 ./sinh(A)

csc(A) Cosecant 1 ./sin(A)

cumprod Cumulative product cumprod

▹ C = cumprod(A): If A is a matrix, cumprod(A) is equivalent to cumprod(A,1) in Matlab whereas in Scilab cumprod(A) gives the cumulative product of all the entries of A taken columnwise. Actually, Matlab works on the first non-singleton dimension and Scilb does not. ▹ C = cumprod(A,dim): Matlab can work with dim greater than number of dimensions of A but Scilab can not, in this can use mtlb_cumprod instead. B = cumprod([1,2,3;4,5,6]) B = [1,2,3;4,10,18] B = cumprod([1,2,3;4,5,6],1) B = [1,2,3;4,10,18] B = cumprod([1,2,3;4,5,6]) B = [1,8,120;4,40,720] B = cumprod([1,2,3;4,5,6],1) B = [1,2,3;4,10,18]

cumsum Cumulative sum cumsum

▹ C=cumsum(A): If A is a matrix, cumsum(A) is equivalent to cumsum(A,1) in Matlab whereas in Scilab cumsum(A) gives the cumulative sum of all the entries of A taken columnwise. Actually, Matlab works on the first non-singleton dimension and Scilb does not. ▹ C = cumsum(A,dim): Matlab can work with dim greater than number of dimensions of A but Scilab can not, in this can use mtlb_cumsum instead. B = cumsum([1,2,3;4,5,6]) B=[1,2,3;5,7,9] B = cumsum([1,2,3;4,5,6],1) B=[1,2,3;5,7,9] B = cumsum([1,2,3;4,5,6]) B=[1,7,15;5,12,21] B = cumsum([1,2,3;4,5,6],1) B=[1,2,3;5,7,9]

	Matlab	D

date Current date string date()

dec2bin The binary representation of a decimal number dec2bin

dec2hex Decimal to hexadecimal number conversion dec2hex

▹ Empty matrix input: In Matlab dec2hex returns "" when input is [] but Scilab dec2hex returns []. ▹ Complex inputs: In Matlab dec2hex automatically removes complex part of input but not in Scilab. ▹ Two inputs: In Matlab dec2hex can have two inputs, in Scilab mtlb_dec2hex emulates this case.

delete Delete files or graphics objects mdelete ↔ delete

▹ Files: When Matlab delete is used to delete a file, Scilab equivalent is mdelete. ▹ Graphics objects: When Matlab delete is used to delete a graphics object, Scilab equivalent is delete. Note that Scilab delete can delete a set of graphics handles is its input is a matrix of such objects.

det Determinant det

diag Diagonal including or extracting diag

Due to the fact that strings or not considered in the same way in Matlab and in Scilab, results are not equal if A is a string matrix or vector in diag(A) or diag(A,k). Note that mtlb_diag can emulate this particularity in Scilab. B = diag('str') B = ['s ';' t ';' r'] B = diag(``str'') B = ``str'' B = mtlb_diag(``str'') B = [``s ``;'' t ``;'' r'']

diary Save session to a file diary

When a filename is given to save environment, if this file exists, Scilab returns an error message while Matlab save environment at the end of existing file (append). Note that diary on and diary toggle exist only in Matlab. The equivalent for Matlab diary off is diary(0) in Scilab.

diff Differences and approximate derivatives diff

▹ Y = diff(X[,n]): For this kind of use of diff (dim parameter not given), Matlab works on the first non-singleton dimension of X what Scilab does not. In this case, Scilab considers dim to be "*" so that diff threats all values of X, what Matlab does not. ▹ Y = diff(X,n,dim): If dimension given by dim reaches 1 before n iterations have been done, Matlab switches to next non-singleton dimension, but Scilab does not, use mtlb_diff to get equivalent results...When n is greater than all existing dimensions of X, Matlab returns [] what Scilab may not give for all inputs...

dir Display directory listing dir

When used in command window, Scilab and Matlab dir are equivalents. When result is stored in a value, Matlab returns a struture but Scilab returns a tlist. To get the same result, you can use mtlb_dir, note that in this case, hidden files are not get.

display Overloaded method to display an object display

disp Display text or array disp

docopt Web browser for UNIX platforms No equivalent

There no Scilab equivalent function, however, information about Web Browser used can be found using globalvariable %browsehelp. Thos variables exists under all platforms.

doc Display online documentation help

dos Execute a UNIX command and return result unix_g

Output values order is inversed in Scilab and in Matlab. In Scilab use disp to emulate Matlab -echo option. [status,result] = dos(...) [result,status] = unix_g(...)

double Conversion to double precision double

In Matlab, this function returns a Boolean type value for a Boolean input whereas Scilab function returns a Double type value.

drawnow Complete pending drawing events No equivalent

In Scilab, drawing events are executed immediately. Scilab drawnow is different from Matlab one.

	Matlab	E

echo Echo lines during execution mode

Scilab mode and Matlab echo are not exactly equivalents but they both change the information displayed during execution. Scilab mode has to be called inside a script or a function but Matlab echo can be called from prompt. However, some uses are equivalents such as: ▹ echo: is equivalent to Scilab mode(abs(mode()-1)) for scripts and non-compiled functions ▹ echo on: is equivalent to Scilab mode(1) for scripts and non-compiled functions ▹ echo off: is equivalent to Scilab mode(0)

eig Find eigenvalues and eigenvectors spec ↔ bdiag

▹ eig(A): Scilab equivalent for eig(A) is spec(A). Scilab eigen vector matrix can differ from Matlab one... ▹ eig(A,'nobalance'): There is no Scilab equivalent for 'nobalance' option. See examples... ▹ eig(A,B,flag): There is no Scilab equivalent for flag. d = eig(A,'balance') [V,D] = eig(A,'balance') d = eig(A,B) [V,D] = eig(A,B) d = spec(A) [V,D] = bdiag(A+%i,1/%eps) [al,be] = spec(A); d = al./be; [al,be,V] = spec(A); D = spec(al./be);

elseif Conditionally execute statements elseif

else Conditionally execute statements else

end Terminate loops and conditionals end

erfcx Scaled complementary error function erfcx

erfc Complementary error function erfc

erf Error function erf

error Display error messages error

Scilab error function can only take one character string as input but Matlab function can take more than one character string as input and also numerical values...

etime Elapsed time etime

In Scilab, etime can be used with 6 and 10 value vectors but Matlab etime can only be used with 6 value vectors ([Year Month Day Hour Minute Second]).

eval Execute a string containing an instruction/expression evstr ↔ execstr

▹ Expression: When eval has to execute an expression then Scilab equivalent for eval is evstr ▹ Instruction: When eval has to execute an instruction with just one output value then Scilab equivalent for eval is evstr. If instruction has more than one output value then execstr has to be used as follows.When eval is used with two inputs then an equivalent can be found in examples below. eval('1+1') eval('x=1+1') eval('[d1,d2]=size(1)') [d1,d2]=eval('size(1)') eval('1+1','1+2') evstr('1+1') x = evstr('1+1') execstr('[d1,d2]=size(1)') execstr('[d1,d2]=size(1)') if execstr("1+1","errcatch") then execstr("1+2");end

exist Check if a variable or file exists exist

Scilab exist function only works for variables, not for M-files or else... Scilab mtlb_exist function is a partial emulation for Matlab exist

exit Ends current session exit

expm Matrix exponential expm

exp Exponential exp

eye Identity matrix eye

Note that in Matlab, A can contain complex values (in these cases, only real part of A is taken in account), what Scilab function do not tolerate. ▹ B=eye(A): If A is a scalar, then Matlab returns a A*A identity matrix but in Scilab you get a 1, use eye(A,A) to get the same matrix B. If A is a vector, Scilab and Matlab give the same B. Finally, if A is a matrix, in Scilab, B will be a matrix having the same size as A whereas in Matlab, you get an error message. B = eye(2) B = [1,0;0,1] B = eye(2,2) B = [1,0;0,1] B = eye([3,3]) B = [1,0,0;0,1,0;0,0,1] B = eye(2) B = 1 B = eye(2,2) B = [1,0;0,1] B = eye([3,3]) B = [1,0]

	Matlab	F

factor Prime numbers decomposition factor

false False array No equivalent

To get the same result in Scilab, use: zeros(...)==1. See zeros.

fclose Close one or more open files mclose

feof Test for end-of-file meof

ferror Query about errors in file input or output mclearerr ↔ merror

▹ ferror(fid): When Matlab ferror is called with just one input and one output, then Scilab equivalent is merror. ▹ ferror(fid,'clear'): When Matlab ferror is called with two inputs and just one output, then Scilab equivalent is mclearerr.For all other cases, there no equivalent in Scilab.

feval Function evaluation evstr ↔ execstr

▹ One output: In this case Scilab evstr is an equivalent to feval, after modifying inputs such as in examples below. ▹ More than one output: In this case Scilab execstr is an equivalent to feval, after modifying inputs such as in examples below. [y1] = feval(@cos,0) [y1,y2] = feval(@size,1) y1 = evstr("cos(0)") execstr("[y1,y2] = size(1)")

fftshift Shift zero-frequency component of discrete Fourier transform to center of spectrum fftshift

▹ fftshift(A[,dim]): Due to the fact that strings or not considered in the same way in Matlab and in Scilab, results are not equal if A is a string matrix or vector in fftshift(A) or fftshift(A,dim). mtlb_fftshift can emulate this particularity in Scilab. ▹ fftshift(A,dim): In Matlab, dim can be greater than the number of dimensions of A but in Scilab you get an error message in this case. mtlb_fftshift can emulate this particularity in Scilab. Y = fftshift('str') Y = 'rst' Y = fftshift('str') Y = 'str' Y = mtlb_fftshift('str') Y = 'rst'

fft(A[,...]) Discrete Fourier transform fft(A,-1[,...])

▹ Y = fft(X): If X is a vector then Scilab equivalent for Matlab fft(X) is fft(X,-1). If X is a matrix then Scilab equivalent for Matlab fft(X) is fft(X,-1,2,1). ▹ Y = fft(X,n) / Y = fft(X,n,dim) / Y = fft(X,[],dim): There is no Scilab equivalent for all these Matlab uses of fft, in these cases, use mtlb_fft instead.

fgetl Read line(s) from file, discard newline character mgetl

Matlab fgetl reads file line per line while Scilab mgetl allows to read the whole file. Matlab fgetl returns -1 if it could not read a line in file but Scilab mgetl returns an empty string is this case. You can used meof to check if End Of File has been reached.

fgets Read line from file, keep newline character fgetstr

Input parameter order is inversed in Scilab compared to Matlab. fgets(fid,n) fgetstr(n,fid)

fileparts Return filename parts fileparts

Scilab function does not get file version but Matlab one does. Scilab function can take a second input parameter specifying the output value we want to get saying: "path", "fname" or "extension".

filesep Return the directory separator for this platform No equivalent

There is no Scilab function equivalent to Matlab filesep but the same output can be obtained with pathconvert("/").

findstr Find one string within another No equivalent

There is no equivalent for findstr in Scilab, but an emulation function has been written: mtlb_findstr. Scilab strindex function is very similar to findstr but do not consider the size of the strings passed as parameters. strindex can replace findstr only if findstr can be replaced by strfind in Matlab.

find Find indices and values of nonzero elements find

Matlab function can work with complex values what is not possible in Scilab, however, using abs it is very easy to have the same results. Note that Scilab function can only return two output values and Matlab one can return a third value that can be computed according to the first two output matrices as explained in Matlab help. For example, in [i,j,v]=find(X), v is equal to: X(i+(j-1))*size(X,1). Another great difference between Scilab and Matlab is that Matlab returns column vectors of indices when X is a column vector or a matrix but Scilab always returns row vectors. For this kind of input, use matrix to get the same output value...what is done mtlb_find()

fix Round towards zero fix

fliplr(A) Flip matrix in left/right direction A(:,$:-1:1)

Due to the fact that Scilab and Matlab do not consider character string in the same way, result is different for this kind of input, in this case, use mtlb_fliplr instead.

flipud(A) Flip matrix in up/down direction A($:-1:1,:)

floor Round down floor

fopen Open a file or obtain information about open files mopen

▹ Access permission: Matlab offers two permissions options not supported by Scilab: W and A (for tape drives) ▹ Input values: In Matlab, fopen('all') lists all opened files, in Scilab, this type of call for fopen does not exist. You can also use fopen in Matlab to get informations on a file identifier (fopen(fid)), this case is not implemented in Scilab. ▹ Machine format: Note that Scilab does not support machine format values as input or output.Matlab fopen can return an error message but not Scilab mopen, moreover, returned file identifier is -1 in case of error in Matlab but does not exist in this case in Scilab, so an emulation function has been written mtlb_fopen.

format Control display format for output format

Some Matlab formats have no Scilab equivalent, see examples below. format format + format bank format compact format hex format long format long e format long g format loose format rat format short format short e format short g format("v",6) format(6) No equivalent for: format "bank" No equivalent for: format "compact" No equivalent for: format "hex" format(16) format("e",16) format("e",16) No equivalent for: format "loose" No equivalent for: format "rat" format(6) format("e",6) format("e",6)

for Repeat statements a specific number of times for

The variable used as loop index is clear in Scilab if all iterations have been made but is not clear if llop is ended by a break. In Matlab, this variable is never cleared.

fprintf Write formatted data to file No equivalent

There is no equivalent function for Matlab fprintf in Scilab but an emulation function has been written: mtlb_fprintf. This function probably not allows all Matlab fprintf possibilities...

fread Read binary data to a file No equivalent

There is no equivalent function for Matlab fread in Scilab but an emulation function has been written: mtlb_fread. This function probably not allows all Matlab fread possibilities (skip parameter is ignored...).

frewind(fid) Move the file position indicator to the beginning of an open file mseek("0",fid)

fscanf Read formatted data to file No equivalent

There is no equivalent function for Matlab fscanf in Scilab but an emulation function has been written: mtlb_fscanf. This function probably not allows all Matlab fscanf possibilities...

fseek Set file position indicator mseek

Scilab and Matlab functions differ by the flag which indicate the origin of the position indicator, see examples below. Note that order of input value is different... File beginning: fseek(fid,offset,'bof') fseek(fid,offset,-1) Current position: fseek(fid,offset,'cof') fseek(fid,offset,0) File end: fseek(fid,offset,'eof') fseek(fid,offset,1) File beginning: fseek(offset,fid,"set") Current position: fseek(offset,fid,"cur") File end: fseek(offset,fid,"end")

ftell Get file position indicator mtell

fullfile Build a full filename from parts fullfile

full Convert sparse matrix to full matrix full

function Function definition function

fwrite Write binary data to a file No equivalent

There is no equivalent function for Matlab fwrite in Scilab but an emulation function has been written: mtlb_fwrite. This function probably not allows all Matlab fwrite possibilities (skip parameter is ignored...).

	Matlab	G

gammaln Logarithm of gamma function gammaln

gamma Gamma function gamma

getenv Get environment variable getenv

Scilab getenv allows to set the string that will be returned if environment variable we want to get does not exist, but not Matlab function.

global Define a global variable global

graymon Set graphics defaults for gray-scale monitors No equivalent

This Matlab function can be replaced in Scilab by: set(gdf(),"color_map",[0.75,0.5,0.25]'*ones(1,3)).

grid Grid lines for two- and three-dimensional plots No equivalent

There is no minor grid in Scilab. There is no equivalent function for Matlab grid function in Scilab but it has equivalents: ▹ grid on: may be replaced by set(gca(),"grid",[1 1]) ▹ grid off: may be replaced by set(gca(),"auto_clear",[-1 -1]) ▹ grid minor: can be emuled in Scilab by mtlb_hold but all grids are toggled ▹ grid: can be emuled in Scilab by mtlb_hold ▹ grid(axes_handle,'on'): may be replaced by axes_handle.grid=[1 1] ▹ grid(axes_handle,'off'): may be replaced by axes_handle.grid=[-1 -1] ▹ grid(axes_handle,'minor'): can be emuled in Scilab by mtlb_hold but all grids are toggled ▹ grid(axes_handle): can be emuled in Scilab by mtlb_hold(axes_handle)

	Matlab	H

hankel Hankel matrix hank

The main difference between Scilab and Matlab function is that they do not use the same input values to build an Hankel matrix. If in Matlab, you just have to give a column vector (and eventually a row vector), Scilab function requires the size of the Hankel matrix to build and a covariance sequence vector for this matrix. (See syntax below) H1 = hankel(C1) H2 = hankel(C2,R2) N1 = size(C1,"*"); COV1 = [matrix(C1,1,-1),zeros(1,N1)]; H1 = hank(N1,N1,COV1); M2 = size(C2,"*"); N2 = size(R2,"*"); COV2 = [matrix(C2,1,-1),matrix(R2(2:$),1,-1)]; H2 = hank(M2,N2,COV2);

helpbrowser Display Help browser for access to full online documentation help

helpdesk Display Help browser help

helpwin Provide access to and display help for all functions help

help Display help help

In Scilab you can not get help on a toolbox, only on a function... Scilab equivalent for Matlab help syntax is help("names").

hess Hessenberg form of a matrix hess

hold Hold current graph No equivalent

There is no equivalent function for Matlab hold function in Scilab but it has equivalents: ▹ hold on: may be replaced by set(gca(),"auto_clear","off") ▹ hold off: may be replaced by set(gca(),"auto_clear","on") ▹ hold: can be emuled in Scilab by mtlb_hold

home Move the cursor to the upper left corner of the Command Window tohome

Note that Matlab function has no effect in no window mode under Unix/Linux while Scilab one clears the terminal display as if you were using "clear" command.

horzcat Horizontal concatenation No equivalent

Scilab equivalent for Matlab horzcat(A1,A2,...,An) is [A1,A2,...,An] if all Ai are not character strings, else, Scilab equivalent is A1+A2+...+An.

	Matlab	I

ifft(A[,...]) Inverse discrete Fourier transform fft(A,1[,...])

▹ Y = ifft(X): If X is a vector then Scilab equivalent for Matlab ifft(X) is fft(X,1). If X is a matrix then Scilab equivalent for Matlab ifft(X) is fft(X,1,2,1). ▹ Y = ifft(X,n) / Y = ifft(X,n,dim) / Y = ifft(X,[],dim): There is no Scilab equivalent for all these Matlab uses of ifft, in these cases, use mtlb_ifft instead.

if Conditionally execute statements if

In Scilab condition can be ended by then but not in Matlab.

imag Complex imaginary part imag

input Request user input input

int16 Convert to 16-bit signed integer int16

For infinite and NaNs values, Scilab and Matlab int16 return different values. A = int16(inf) A = 32767 A = int16(-inf) A = -32768 A = int16(nan) A = 0 A = int16(%inf) A = -32768 A = int16(-%inf) A = -32768 A = int16(%nan) A = -32768

int32 Convert to 32-bit signed integer int32

For infinite and NaNs values, Scilab and Matlab int32 return different values. A = int32(inf) A = 2147483647 A = int32(-inf) A = -2147483648 A = int32(nan) A = 0 A = int32(%inf) A = -2147483648 A = int32(-%inf) A = -2147483648 A = int32(%nan) A = -2147483648

int8 Convert to 8-bit signed integer int8

For infinite values, Scilab and Matlab int8 return different values. A = int8(inf) A = 127 A = int8(-inf) A = -128 A = int8(%inf) A = 0 A = int8(-%inf) A = 0

interp1 One_dimension interpolation function interp1

Scilab interp1 function doesn't work with the pchip interpolation method.

inv Matrix inverse inv

isa Detect an object of a given type No equivalent

There is no equivalent function for Matlab isa function in Scilab but it can be replaced by equivalent syntaxes as shown is examples. a = isa(x,'logical') b = isa(x,'char') c = isa(x,'numeric') d = isa(x,'int8') e = isa(x,'uint8') f = isa(x,'int16') g = isa(x,'uint16') h = isa(x,'int32') k = isa(x,'uint32') l = isa(x,'single') m = isa(x,'double') n = isa(x,'cell') o = isa(x,'struct') p = isa(x,'function_handle') q = isa(x,'sparse') r = isa(x,'lti') a = type(x)==4; b = type(x)==10; c = or(type(x)==[1,5,8]); d = typeof(x)=="int8"; e = typeof(x)=="uint8"; f = typeof(x)=="int16"; g = typeof(x)=="uint16"; h = typeof(x)=="int32"; k = typeof(x)=="uint32"; l = type(x)==1; m = type(x)==1; n = typeof(x)=="ce"; o = typeof(x)=="st"; p = type(x)==13; q = type(x)==5; r = typeof(x)=="state-space";

iscell(A) Determine if input is a cell array typeof(A)=="ce"

ischar(A) Determine if item is a character array type(A)==10

isdir Determine if item is a directory isdir

isempty True for empty matrix isempty

isequal Determine if arrays are numerically equal isequal

In Scilab, struct fields must be in the same order so that structs can be equal but not in Matlab.

isfield Determine if input is a structure array field No equivalent

There is no Scilab equivalent function for Matlab tf=isfield(A,field) but there are equivalent instructions: ▹ If A is not a structure and/or field is not a character string: Scilab equivalent is %F. ▹ If A is a structure and field is a character string: Scilab equivalent is allfields=getfield(1,A);tf=or(allfields(3:$)==field);.

isfinite True for finite elements No equivalent

There is no equivalent function for Matlab isfinite function in Scilab but it can be emuled by: abs(A)<%Inf

isglobal Determine if item is a global variable isglobal

ishandle(A) Determines if values are valid graphics object handles type(A)==9

ishold Return hold state No equivalent

There is no equivalent function for Matlab ishold function in Scilab but it can be emuled by: get(gca(),"auto_clear")=="off";.

isinf True for infinite elements isinf

isinteger(A) Detect whether an array has integer data type type(A)==8

isletter True for letters of the alphabet No equivalent

There is no equivalent function to Matlab isletter function in Scilab. However it can be replaced as follows. Using mtlb_isletter will give a prettier code. tf = isletter(A) If A is a String matrix: tf = ((asciimat(A)$>=$65) & (asciimat(A)90)) $|$ ((asciimat(A)$>=$97) & (asciimat(A)122)) If A is not a String matrix: tf = zeros(A)

islogical(A) Determine if item is a logical array No equivalent

There is no equivalent function for Matlab islogical function in Scilab but it can be emuled by: or(type(A)==[4,6]).

isnan Detect NaN elements of an array isnan

isnumeric(A) Determine if input is a numeric array No equivalent

There is no equivalent function for Matlab isnumeric function in Scilab but it can be emuled by: or(type(A)==[1 5 8]).

ispc Determine if PC (Windows) version MSDOS

isreal Determine if all array elements are real numbers isreal

Scilab isreal function can take two values as input. The first one is the same as Matlab one and the second allows to give a tolerance on the absolute value of the imaginary part of first input. So to have the same results in Matlab and in Scilab, second input in Scilab function must be set to 0. tf = isreal(1+0i) tf = 1 tf = isreal(1+0*%i) tf = %F tf = isreal(1+0*%i,0) tf = %T

isscalar(A) Determine if input is scalar sum(length(A))==1

isspace Detect elements that are ASCII white spaces No equivalent

There is no equivalent function to Matlab isspace function in Scilab. However it can be replaced as shown below. tf = isspace(A) If A is a String matrix: tf = asciimat(A)==32 If A is not a String matrix: tf = zeros(A)

issparse(S) Test if matrix is sparse No equivalent

There is no equivalent function for Matlab issparse function in Scilab but it can be emuled by: or(type(S)==[5,6]).

isstruct(A) Determine if input is a structure array typeof(A)=="st"

isstr(A) Determine if item is a character array type(A)==10

isunix Determine if Unix version ~MSDOS

isvector Determine if input is a vector No equivalent

There is no Scilab equivalent function for Matlab tf=isvector(A) but there are equivalent instructions: ▹ If A is not a character string: Scilab equivalent is tf = size(A,1)==1 | size(A,2)==1. ▹ If A is a character string: Scilab equivalent is tf = size(asciimat(A),1)==1 | size(asciimat(A),2)==1.

	Matlab	K

keyboard Invoke the keyboard in a file pause

kron(A,B) Kronecker tensor product A .*. B

	Matlab	L

length(A) Length of vector No equivalent

If A is a vector, Scilab equivalent for length(A) is size(A,"*"). If A is a matrix, Scilab equivalent for length(A) is max(size(A)). If A contains character strings, String matrix has to be converted to a "character" string matrix using mstr2sci (Using asciimat to convert will give the same result). Scilab length is different from Matlab length.

linspace Linearly spaced vector linspace

When A and/or B is a String in Matlab, linspace(A,B[,n]) returns a String matrix, in Scilab, it can be made with ascii(linspace(ascii(A),ascii(B),n)).

load Load workspace variables from disk loadmatfile

Scilab loadmatfile does not handle option -regexp yet.

log10 Common (base 10) logarithm log10

log2 Base 2 logarithm and dissect floating point number log2 ↔ frexp

Scilab log2 is equivalent to Matlab log2 for logarithm computation, but for floating point number dissection, Scilab equivalent to Matlab log2 is frexp.

logical(A) Convert numeric values to logical No equivalent

If A is not an empty matrix, Scilab equivalent for logical(A) is is not equal to 0 else Scilab equivalent is [].

log Natural logarithm log

lookfor Search for specified keyword in all help entries apropos

No Scilab equivalent for Matlab -all option.

lower(str) Convert string to lower case convstr(str,''u'')

If A is not a character string matrix, Scilab equivalent for B=lower(A) is B=A, else equivalent is B=convstr(A).

lu LU matrix factorization lu

There is no Scilab equivalent for Matlab lu when called with 1 or 4 outputs or with 2 inputs.

	Matlab	M

max Maximum max

Matlab max function can work with complexes, what Scilab max can not, so a emulation function called mtlb_max has been written. Note that in Scilab, second input parameter can give the dimension to use to find the maximum values or another matrix (maximum of two matrices), in Matlab, dimension parameter is given in a third input parameter (in this case, second parameter must be []). ▹ C=max(A): If A is a matrix, max(A) is equivalent to max(A,[],1) in Matlab whereas in Scilab max(A) gives the maximum value found in A. Matlab max treats the values along the first non-singleton dimension. A = [1,2,3;4,5,6] C = max(A) C = [4,5,6] C = max(A,[],1) C = [4,5,6] B=[7,8,9;2,3,4] C = max(A,B) C = [7,8,9;4,5,6] A = [1,2,3;4,5,6] C = max(A) C = 6 C = max(A,''r'') C = [4,5,6] B=[7,8,9;2,3,4] C = max(A,B) C = [7,8,9;4,5,6]

min Minimum min

Matlab min function can work with complexes, what Scilab min can not, so a emulation function called mtlb_min has been written. Note that in Scilab, second input parameter can give the dimension to use to find the minimum values or another matrix (minimum of two matrices), in Matlab, dimension parameter is given in a third input parameter (in this case, second parameter must be []). ▹ C=min(A): If A is a matrix, min(A) is equivalent to min(A,[],1) in Matlab whereas in Scilab min(A) gives the minimum value found in A. Matlab min treats the values along the first non-singleton dimension. A = [1,2,3;4,5,6] C = min(A) C = [1,2,3] C = min(A,[],1) C = [1,2,3] B = [7,8,9;2,3,4] C = min(A,B) C = [1,2,3;2,3,4] A = [1,2,3;4,5,6] C = min(A) C = 1 C = min(A,''r'') C = [1,2,3] B = [7,8,9;2,3,4] C = min(A,B) C = [1,2,3;2,3,4]

mkdir mkdir

Scilab mkdir returns 1 or 2 values but Matlab one can return up to three values (third output is a Matlab messageid).

mod Modulus after division pmodulo

Scilab pmodulo can work with Complex values what Matlab mod can not.

more Display Command Window output one screenful at a time lines

See examples. more off more on more(30) lines(0) lines(60) lines(30)

	Matlab	N

nargin ↔ nargin('fun') Number of function input arguments argn(2) ↔ size(getfield(1,macrovar(fun)),"*")

nargout ↔ nargout('fun') Number of function output arguments argn(1) ↔ size(getfield(2,macrovar(fun)),"*")

ndims Number of array dimensions No equivalent

There is no Scilab equivalent function for ndims(A) but it can be emuled by: size(size(A),"*")

norm Vector and matrix norms norm

num2str Number to string conversion string ↔ msprintf

▹ num2str(a,precision): There is no Scilab equivalent for this Matlab expression. ▹ num2str(a,format): Scilab equivalent for Matlab num2str(a,format) is msprintf(format,a).

	Matlab	O

ones Create an array of all ones ones

Note that in Matlab, A can contain complex values (in these cases, only real part of A is taken in account), what Scilab function do not tolerate. ▹ B=ones(A): If A is a scalar, then Matlab returns a A*A matrix of ones but in Scilab you get a 1, use ones(A,A) to get the same matrix B. If A is a vector, Scilab and Matlab give the same B. Finally, if A is a matrix, in Scilab, B will be a matrix having the same size as A whereas in Matlab, you get an error message. B = ones(2) B = [1,1;1,1] B = ones(2,2) B = [1,1;1,1] B = ones([3,3]) B = [1,1,1;1,1,1;1,1,1] B = ones(2) B = 1 B = ones(2,2) B = [1,1;1,1] B = ones([3,3]) B = [1,1]

otherwise Default part of switch/select statement else

	Matlab	P

pause Halt execution temporarily xpause ↔ halt

Scilab equivalent for Matlab pause(n) is xpause(1000*n). ▹ pause ↔ halt(): Scilab halt() and Matlab pause are equivalents. ▹ pause on/off: There is no Scilab equivalent for Matlab pause on or pause off

perms Array of all permutations of vector components perms

permute Permute the dimensions of an array permute

pie circular graphic pie

plot Linear 2-D plot plot

Scilab plot doesn't accept all the properties of the Matlab plot

pow2 Base 2 power and scale floating-point numbers No equivalent

▹ X=pow2(Y): There is not equivalent function for pow2 in Scilab but, when called with one input argument it can be emulated by: 2^ Y ▹ X=pow2(F,E): In this case, Matlab pow2() ignores imaginary part of input arguments. An equivalent expression for this use of pow2 is: F.* 2 .^ E (Note that 2 must be preceeded and followed by a white space).

primes Returns the primes numbers included between 1 and given number primes

prod Product of array elements prod

▹ M=prod(A): Scilab prod(A) returns the product of all components of A. So, if A is a vector, then Scilab and Matlab work in the same way. If A is a matrix, Scilab prod(A) gives the product of all elements of A but Matlab returns the product of each column. Finally, if A is a multidimensional array, Matlab works on the first non-singleton dimension of A what Scilab does not. So, to be sure to find a Scilab equivalent for Matlab call to prod it is better to precise dimension on which to work. ▹ M=prod(A,dim): In Scilab dim=1 is equivalent to dim=''r'' and dim=2 is equivalent dim=''c''. In Matlab, dim can be greater then the number of dimension of A (in this case, M=A), in Scilab you will get an error message. A = [1,2,3;4,5,6] M = prod(A) M = [4,10,18] M = prod(A,1) M = [4,10,18] A = [1,2,3;4,5,6] M = prod(A) M = 720 M = prod(A,''r'') M = [4,10,18]

	Matlab	Q

qr Orthogonal-triangular decomposition qr

When used with two input values and tree output values, Scilab and Matlab qr results can differ. Use mtlb_qr instead.

quit Terminate session quit

	Matlab	R

randn(A) Normally distributed random numbers and arrays rand(A,''normal'')

▹ B=randn(A) ↔ B=rand(A,``normal''): If A is a scalar, then Matlab returns a A*A random matrix but in Scilab you get a single random value, use rand(A,A,''normal'') to get the same matrix B. Finally, if A is a matrix, in Scilab, B will be a matrix having the same size as A whereas in Matlab, you get an error message.Note that in Matlab, A can contain complex values (in these cases, only real part of A is taken in account), what Scilab function do not tolerate. Particular case: To get the state of the normal generator, in Matlab you have to use s=randn('state') to get 2 current values of the generator, but Scilab equivalent s=rand(``seed'') return only one value.

rand(A) Uniformly distributed random numbers and arrays rand(A[,''uniform''])

▹ B=rand(A): If A is a scalar, then Matlab returns a A*A random matrix but in Scilab you get a single random value, use rand(A,A) to get the same matrix B. Finally, if A is a matrix, in Scilab, B will be a matrix having the same size as A whereas in Matlab, you get an error message.Note that in Matlab, A can contain complex values (in these cases, only real part of A is taken in account), what Scilab function do not tolerate. Particular case: To get the state of the uniform generator, in Matlab you have to use s=rand('state') to get 35 current values of the generator, but Scilab equivalent s=rand(``seed'') return only one value.

rcond Matrix reciprocal condition number estimate rcond

Scilab and Matlab values differ for empty matrix. c = rcond([]) c = Inf c = rcond([]) c = []

realmax Largest positive floating-point number number_properties("huge")

There is no Scilab equivalent for Matlab realmax('single').

realmin Smallest positive floating-point number number_properties("tiny")

There is no Scilab equivalent for Matlab realmin('single').

real Real part of a complex number real

rem(X,Y) Remainder after division X-fix(X./Y).*Y

repmat Replicate and tile an array No equivalent

There is no Scilab equivalent function for Matlab repmat but it can be replaced by the following expressions (considering m and n being real values): ▹ repmat(A,m) with m a scalar: can be replaced by ones(m,m).*.A if A is of Double type, by ones(m,m).*.bool2s(A) if A is of Boolean type and by asciimat(ones(m,m).*.asciimat(A) if A is of String type ▹ repmat(A,m) with m a vector: can be replaced by ones(m(1),m(2),...).*.A is of Double type, by ones(m(1),m(2),...).*.bool2s(A) if A is of Boolean type and by asciimat(ones(m(1),m(2),...).*.asciimat(A) if A is of String type ▹ repmat(A,m,n): can be replaced by ones(m,n).*.A if A is of Double type, by ones(m,n).*.bool2s(A) if A is of Boolean type and by asciimat(ones(m,n).*.asciimat(A) if A is of String type

reshape Reshape array matrix

To get the same result for character string matrices in Scilab than in Matlab, convert Scilab character string matrices using mstr2sci. All unspecified dimensions are represented by a [] input in Matlab while in Scilab they are given by a -1. Matlab reshape suppresses singleton higher dimension, it is not the case for matrix in Scilab...

return Return to the invoking function return

round Round to nearest integer round

	Matlab	S

save Save workspace variables from disk mtlb_save

Scilab mtlb_save does not handle options -v4 -mat and -append yet.

schur Schur decomposition schur

setstr Set string flag ascii

▹ S = setstr(A) with A a caracter string: In this case, Scilab ascii function convert string to ascii code matrix, but setstr keeps string format.

sign Signum function sign

sinh Hyperbolic sine sinh

sin Sine sin

size Array dimension size

Due to the fact that strings or not considered in the same way in Matlab and in Scilab, results are not equal for string matrices, convert it using m2scistr to have the same result. ▹ d = size(X,dim): If dim is greater than number of dimensions of X, Matlab returns d=1, but in Scilab, you get an error message. Scilab mtlb_size can work with dim greater than number of dimensions of X. ▹ [d1,...dn] = size(X): If n is greater than number of dimensions of X, all "extra" variables are set to 1 in Matlab but Scilab returns an error message. Scilab mtlb_size returns a Matlab like result in these cases. When n is less than number of dimensions of X, dn contains the product of the sizes of the remaining dimensions in Matlab but in Scilab dn = size(X,n), use mtlb_size for such uses.

sort Sort elements in ascending order No equivalent

Scilab sort and Matlab sort are different functions ! For character string inputs, please use better mtlb_sort in Scilab... ▹ B = sort(A): Scilab gsort can be used as an equivalent for Matlb sort giving it the good inputs. If A is a vector, call gsort(A,"g","i"). If A is a matrix then call gsort(A,"r","i"). Note that gsort does not work with multidimensional arrays... ▹ B = sort(A,dim): If in Matlab, dim is 1 (respectively 2) then replace it by "r" (respectively "c") in Scilab when calling gsort(A,dim,"i"). Note that gsort does not work with multidimensional arrays...

sparse Create sparse matrix sparse

Matlab and Scilab equivalents: ▹ sparse(A) ↔ sparse(A): ▹ sparse(m,n) ↔ sparse([],[],[m,n]): ▹ sparse(i,j,s) ↔ sparse([i,j],s): This equivalence is true considering i, j and s have the same length and that i and j are column vectors. ▹ sparse(i,j,s,m,n) ↔ sparse([i,j],s,[m,n]): This equivalence is true considering i, j and s have the same length and that i and j are column vectors. ▹ sparse(i,j,s,m,n,nzmax): There is no Scilab equivalent for this use of Matlab sparse.

sqrt Square root sqrt

strcmpi(str1,str2) Compare strings ignoring case convstr(str1)==convstr(str2)

Note that strcmpi can be use with not string inputs, in this case Matlab returns 0. Scilab == will in this case return %T if both inputs are equal...

strcmp(str1,str2) Compare strings str1==str2

Note that strcmp can be use with not string inputs, in this case Matlab returns 0. Scilab == will in this case return %T if both inputs are equal...

strfind Find one string within another strindex

Note that strfind can be use with not string inputs, in this case Matlab returns 1 if inputs are equal and 0 else but strindex can not do such comparison...

strrep String search and replace strsubst

Note that Matlab strrep can be use with not string inputs, what Scilab strsubst can not. In this case use mtlb_strrep instead.

struct Create struct array struct

sum Sum of array elements sum

▹ M=sum(A): Scilab sum(A) returns the sum of all components of A. So, if A is a vector, then Scilab and Matlab work in the same way. If A is a matrix, Scilab sum(A) gives the sum of all elements of A but Matlab returns the sum of each column. Finally, if A is a multidimensional array, Matlab works on the first non-singleton dimension of A what Scilab does not. So, to be sure to find a Scilab equivalent for Matlab call to sum it is better to precise dimension on which to work. ▹ M=sum(A,dim): In Scilab dim=1 is equivalent to dim=''r'' and dim=2 is equivalent dim=''c''. In Matlab, dim can be greater then the number of dimension of A (in this case, M=A), in Scilab you will get an error message. A = [1,2,3;4,5,6] M = sum(A) M = [5,7,9] M = sum(A,1) M = [5,7,9] A = [1,2,3;4,5,6] M = sum(A) M = 21 M = sum(A,''r'') M = [5,7,9]

surf 3-D surface plot surf

Scilab surf doesn't accept all the properties of the Matlab surf

svd Singular value decomposition svd

switch Switch among several cases based on expression select

	Matlab	T

tanh Hyperbolic tangent tanh

tan Tangent tan

tic Starts a stopwatch timer tic()

In Scilab, tic can be called as a command when output value is just displayed.

title Display a title on a graphic window title

toc Read the stopwatch timer toc()

In Scilab, toc can be called as a command when output value is just displayed.

toeplitz Toeplitz matrix toeplitz

toeplitz can be used with empty matrices in Scilab but not in Matlab. ▹ T=toeplitz(c): If c is complex, use mtlb_toeplitz in Scilab to have the same result than Matlab.Else if c is not a scalar or a vector, use mtlb_toeplitz in Scilab to have the same result than Matlab. ▹ T=toeplitz(c,r): If c and r are not scalars or vectors or if c(1,1)<>r(1,1), use: mtlb_toeplitz in Scilab to have the same result than Matlab.

tril Lower triangular part of a matrix tril

In L=tril(X) and L=tril(X,k), Scilab function gives different results from Matlab one if X is a String matrix. In this case use mtlb_tril instead. Note that k can be complex in Matlab, in this case, only real part of k is taken in account, Scilab gives an error message for a such use.

triu Upper triangular part of a matrix triu

In U=triu(X) and U=triu(X,k), Scilab function gives different results from Matlab one if X is a String matrix. In this case use mtlb_triu instead. Note that k can be complex in Matlab, in this case, only real part of k is taken in account, Scilab gives an error message for a such use.

true True array No equivalent

To get the same result in Scilab, use: ones(...)==1. See ones.

type List file No equivalent

Scilab mtlb_type is a partial emulation of Matlab type function. Scilab type function does not match with Matlab type !

	Matlab	U

uigetdir Standard dialog box for selecting a directory tk_getdir

uint16 Convert to 16-bit unsigned integer uint16

For infinite values, Scilab and Matlab uint16 return different values. A = uint16(inf) A = 65535 A = uint16(%inf) A = 0

uint32 Convert to 32-bit unsigned integer uint32

For infinite values, Scilab and Matlab uint32 return different values. A = uint32(inf) A = 4294967295 A = uint32(%inf) A = 0

uint8 Convert to 8-bit unsigned integer uint8

For infinite values, Scilab and Matlab uint8 return different values. A = uint8(inf) A = 255 A = uint8(%inf) A = 0

unix Execute a UNIX command and return result unix_g

Output values order is inversed in Scilab and in Matlab. In Scilab use disp to emulate Matlab -echo option. [status,result] = unix(...) [result,status] = unix_g(...)

upper(str) Convert string to upper case convstr(str,''u'')

If A is not a character string matrix, Scilab equivalent for B=upper(A) is B=A, else equivalent is B=convstr(A,''u'').

	Matlab	V

varargin Pass variable numbers of arguments varargin

In Matlab varargin is a cell and in Scilab it is a list.

varargout Return variable numbers of arguments varargout

In Matlab varargout is a cell and in Scilab it is a list.

vertcat Vertical concatenation No equivalent

Scilab equivalent for Matlab vertcat(A1,A2,...,An) is [A1;A2;...;An].

	Matlab	W

waitforbuttonpress Wait for key or mouse button press No equivalent

There is no equivalent function for Matlab w=waitforbuttonpress in Scilab however it can be replaced by: [%v0,%v1,%v2,%v3,%v4] = xclick();w = bool2s(%v0>64);

warning Display warning messages warning

Scilab warning function can only take one character string as input but Matlab function can take more than one character string as input and also numerical values...

while Repeat statements an indefinite number of times while

whos List variables in the workspace whos

Scilab whos is an equivalent for Matlab whos just when called as follows: whos or whos("global")

who List variables in the workspace who

Scilab who is an equivalent for Matlab who just when called as follows: who or who("global")

winqueryreg Get item from Microsoft Windows registry winqueryreg

Scilab function returns a matrix of strings or a int32 value but Matlab function returns a Cell of strings or a int32 value.

	Matlab	X

xlabel Display a string along the x axis xlabel

	Matlab	Y

ylabel Display a string along the y axis ylabel

	Matlab	Z

zeros Create an array of all zeros zeros

▹ B=zeros(A): If A is a scalar, then Matlab returns a A*A matrix of zeros but in Scilab you get a 1, use zeros(A,A) to get the same matrix B. If A is a row vector, Scilab and Matlab give the same B. Finally, if A is a matrix, in Scilab, B will be a matrix having the same size as A whereas in Matlab, you get an error message.Note that in Matlab, A can contain complex values (in these cases, only real part of A is taken in account), what Scilab function do not tolerate. B = zeros(2) B = [0,;0,0] B = zeros(2,2) B = [0,0;0,0] B = zeros([3,3]) B = [0,0,0;0,0,0;0,0,0] B = zeros(2) B = 0 B = zeros(2,2) B = [0,0;0,0] B = zeros([3,3]) B = [0,0]

zlabel Display a string along the z axis zlabel

	Matlab	Operators

: Colon :

▹ Using colon with empty matrices: In Matlab if almost one operand is an empty matrix, then result is an empty matrix what gives an error message in Scilab. ▹ Using colon with NaNs and Infs: In Matlab if almost one operand is an empty matrix, then result is a NaN what make Scilab returning an error.

+ Plus +

▹ Character strings addition: In Scilab, string addition is the same as string concatenation, what is done in Matlab by strcat function. In Matlab, string addition is the equivalent of the addition of corresponding ASCII codes. ▹ Empty matrix and addition: In Matlab, addition can only be made if the operands have the same size unless one is a scalar. For exemple, empty matrices can only be added to another empty matrix or a scalar. Note that when you add a scalar and an empty matrix, Matlab result is always an empty matrix while in Scilab, result is equal to the scalar. ▹ Unary plus: In Matlab, unary plus exists, but in Scilab it is automatically deleted when compiling so we can consider that Scilab unary plus does not exist. str = 'str1'+'str2' str = [230,232,228,99] str = strcat('str1','str2') str = 'str1str2' A = 1 + [] A = [] str = 'str1'+'str2' str = 'str1str2' str = strcat(['str1','str2']) str = 'str1str2' A = 1 + [] A = 1

- Minus -

▹ Empty matrix and substraction: In Matlab, substraction can only be made if the operands have the same size unless one is a scalar. For exemple, empty matrices can only be substracted to another empty matrix or a scalar. Note that when you substract an empty matrix to a scalar and inversely, Matlab result is always an empty matrix while in Scilab, result is equal to the scalar. A = 1 - [] A = [] A = 1 - [] A = 1

* Mutiplication *

/ Right division /

\ Left division \

Note that Matlab left division gives strange results when one operand is a character string matrix and not the other one...

== Equal to ==

.* Elementwise mutiplication .*

WARNING: Expressions like X.*.23 are interpreted in Matlab as X elementwisely multiplied by 0.23 while Scilab computes the Kronecker product of X and 23, to have the same result, insert a blank between * and .23

./ Elementwise right division ./

WARNING: Expressions like X./.23 are interpreted in Matlab as the elementwise right division of X by 0.23 while Scilab computes the Kronecker right division of X and 23, to have the same result, insert a blank between / and .23

.\ Elementwise left division .\

WARNING: Expressions like X.\.23 are interpreted in Matlab as the elementwise division of 0.23 by X while Scilab computes the Kronecker left division of X and 23, to have the same result, inser a blank between \ and .23

.' Elementwise transpose .'

▹ Character string elementwise transpose: In Scilab, the result of a character string elementwise transpose is the string itself; but in Matlab, elementwise transpose of a character string gives a column vector of characters. To have the same result in Scilab, use: mtlb_0. s = ('str1')' s = ['s';'t';'r';'1'] s = ('str1')' s = 'str1' s = mtlb_0('str1') s = ['s';'t';'r';'1']

.^ Elementwise exponent .^

Note that Matlab seems to have a bug when exposant is a character... WARNING: Expressions like X.^.23 are interpreted in Matlab as X to the power of 0.23 while Scilab executes X elementwisely powered to 23, to have the same result, inser a blank between ^ and .23

' Transpose '

▹ Character string transpose: In Scilab, the result of a character string transpose is the string itself; but in Matlab, transpose of a character string gives a column vector of characters. To have the same result in Scilab, use: mtlb_t. s = ('str1')' s = ['s';'t';'r';'1'] s = ('str1')' s = 'str1' s = mtlb_t('str1') s = ['s';'t';'r';'1']

& Logical AND &

Due to the fact that strings or not considered in the same way in Matlab and in Scilab, results are not equal for string matrices, convert it to ascii code matrices using m2scistr to have the same result. Scilab function has a bug!

| Logical OR |

Due to the fact that strings or not considered in the same way in Matlab and in Scilab, results are not equal for string matrices, convert it to ASCII code matrices using m2scistr to have the same result. Scilab function has a bug!

> Greater than >

When both operands are empty matrices, Matlab result is an empty matrix while in Scilab you get an error. In Scilab this operator does not work with complex values while in Matlab it considers Real part of them for comparison.

>= Greater or equal to >=

When both operands are empty matrices, Matlab result is an empty matrix while in Scilab you get an error. In Scilab this operator does not work with complex values while in Matlab it considers Real part of them for comparison.

< Smaller than <

When both operands are empty matrices, Matlab result is an empty matrix while in Scilab you get an error. In Scilab this operator does not work with complex values while in Matlab it considers Real part of them for comparison.

<= Smaller or equal to <=

When both operands are empty matrices, Matlab result is an empty matrix while in Scilab you get an error. In Scilab this operator does not work with complex values while in Matlab it considers Real part of them for comparison.

^ Exponent ^

Note that Matlab seems to have a bug for X^(Y) when X is a character and that in Scilab operations such as X^(Y) with X a scalar and Y a matrix is equivalent to X.^(Y) (Will change in next Scilab versions...).

~ Negation ~

Due to the fact that strings or not considered in the same way in Matlab and in Scilab, results are not equal for string matrices, convert it to ascii code matrices using m2scistr to have the same result.

~= Not equal to ~=

	Matlab	Variables

ans The most recent answer ans

end (index) Last index $

eps Floating-point relative accuracy %eps

Only Matlab allows to change the value of this variable and clear eps allows to set the value of eps to its initial value.

i Imaginary unit %i

Only Matlab allows to change the value of this variable.

j Imaginary unit %i

Only Matlab allows to change the value of this variable.

pi Ratio of a circle's circumference to its diameter %pi

Only Matlab allows to change the value of this variable.