Matlab Cheat Sheet

Matlab Cheat Sheet Some nifty commands clc clear clear x commandwindow whos whos x ans close all close(H) winopen(pwd) class(obj) save filename save filename x,y save -append filename x load filename ver beep doc function docsearch string web google.com inputdlg methods(A) Statistical commands Clear command window Clear system memory Clear x from memory open/select commandwindow lists data structures size, bytes, class and attributes of x Last result closes all figures closes figure H Open current folder returns objects class saves all variables to .mat file saves x,y variables to .mat file appends x to .mat file loads all variables from .mat file Lists version and toolboxes Makes the beep sound Help/documentation for function search documentation opens webadress Input dialog box list class methods for A distrnd random numbers from dist distpdf pdf from dist distcdf cdf dist distrnd random numbers from dist hist(x) histogram of x histfit(x) histogram and *Standard distributions (dist): norm, t, f, gam, chi2, bino *Standard functions: mean,median,var,cov(x,y),corr(x,y), *quantile(x,p) is not textbook version. (It uses interpolation for missing quantiles. Built in functions/constants abs(x) absolute value pi 3.1415... inf ∞ eps floating point accuracy 106 1e6 sum(x) sums elements in x cumsum(x) Cummulative sum prod Product of array elements cumprod(x) cummulative product diff Difference of elements round/ceil/fix/floor Standard functions.. *Standard functions: sqrt, log, exp, max, min, Bessel *Factorial(x) is only precise for x < 21 Tables T=table(var1,var2,...,varN) Makes table* T(rows,vars) get sub-table T{rows,vars} get data from table T.var or T.(varindex) all rows of var T.var(rows) get values of var from rows summary(T) summary of table T.var3(T.var3>5)=5 changes some values T.Properties.Varnames Variable names T = array2table(A) ! make table from array T = innerjoin(T1,T2) innerjoin T = outerjoin(T1,T2) outerjoin ! Rows and vars indicate rows and variables. tables are great for large datasets, because they use less memory and allow faster operations. *rowfun is great for tables, much faster than eg. looping Cell commands A cell can contain any variable type. x=cell(a,b) x{n,m} cell2mat(x) cellfun(’fname’,C) a ×b cell array access cell n,m cellfun transforms cell to matrix Applies fname to cells in C Strings and regular expressions strcomp strcompi strncomp strfind regexp compare strings (case sensitive) compare strings (not case sensitive) as strcomp, but only n first letters find string within a string , gives start position Search for regular expression Logical operators && Short-Circuit AND. & AND || Short-Circuit or | or ~ not == Equality comparison ~= not equal isa(obj, ’class_name’) is object in class *Other logical operators: <,>,>=,<= Keyboard shortcuts *All above operators are elementwise edit filename Opens filename in editor *Class indicators: isnan, isequal, ischar, isinf, isvector Alt Displays hotkeys , isempty, isscalar, iscolumn F1 Help/documentation for highlighted function *Short circuits only evaluate second criteria if F5 Run code first criteria is passed, it is therefore faster. F9 Run highlighted code And useful fpr avoiding errors occuring in second criteria F10 Run code line *non-SC are bugged and short circuit anyway F11 Run code line, enter functions Shift+F5 Leave debugger F12 Insert break point Variable generation Ctrl+Page up/down Moves between tabs j:k row vector [j,j+1,...,k] Ctrl+shift Moves between components j:i:k row vector [j,j+i,...,k], Ctrl+C Interrupts code linspace(a,b,n) n points linearly spaced Ctrl+D Open highlighted codes file and including a and b Ctrl+ R/T Comment/uncomment line NaN(a,b) a×b matrix of NaN values Ctrl+N New script ones(a,b) a×b matrix of 1 values Ctrl+W Close script zeros(a,b) a×b matrix of 0 values Ctrl+shift+d Docks window meshgrid(x,y) 2d grid of x and y vectors Ctrl+shift+u Undocks window [a,b]=deal(NaN(5,5)) declares a and b Ctrl+shift+m max window/restore size global x gives x global scope matrix and vector operations/functions x=[1, 2, 3] 1x3 (Row) vector x=[1; 2; 3] 3x1 (Column) vector x=[1, 2; 3, 4] 2x2 matrix x(2)=4 change index value nr 2 x(:) All elements of x (same as x) x(j:end) j’th to last element of x x(2:5) 2nd to 5th element of x x(j,:) all j row elements x(:,j) all j column elements diag(x) diagonal elements of x x.*y Element by element multiplication x./y Element by element division x+y Element by element addition x-y Element by element subtraction A^n normal/Matrix power of A A.^n Elementwise power of A A’ Transpose inv(A) Inverse of matrix size(x) Rows and Columns eye(n) Identity matrix sort(A) sorts vector from smallest to largest eig(A) Eigenvalues and eigenvectors numel(A) number of array elements x(x>5)=0 change elemnts >5 to 0 x(x>5) list elements >5 find(A>5) Indices of elements >5 find(isnan(A)) Indices of NaN elements [A,B] concatenates horizontally [A;B] concatenates vertically For functions on matrices, see bsxfun,arrayfun or repmat *if arrayfun/bsxfun is passed a gpuArray, it runs on GPU. *Standard operations: rank,rref,kron,chol *Inverse of matrix inv(A) should almost never be used, use RREF through \ instead: inv(A)b = A\b. Plotting commands Nonlinear nummerical methods quad(fun,a,b) simpson integration of @fun 2d line plot, handle set to fig1 from a to b change line width fminsearch(fun,x0) minimum of unconstrained dot markers (see *) multivariable function marker type (see *) using derivative-free method line color (see *) fmincon minimum of constrained function marker size (see *) Example: Constrained log-likelihood maximization, note the fonts to size 14 Parms_est = fmincon(@(Parms) -flogL(Parms,x1,x2,x3,y) new figure window ,InitialGuess,[],[],[],[],LwrBound,UprBound,[]); graphics object j returns information graphics object j Debbuging etc. gcf(j) get current figure handle keyboard Pauses exceution subplot(a,b,c) Used for multiple return resumes exceution figures in single plot tic starts timer xlabel(’\mu line’,’FontSize’,14) names x/y/z axis toc stops timer ylim([a b]) Sets y/x axis limits profile on starts profiler for plot to a-b profile viewer Lets you see profiler output title(’name’,’fontsize’,22) names plot try/catch Great for finding where grid on/off; Adds grid to plot errors occur legend(’x’,’y’,’Location’,’Best’) adds legends dbstop if error stops at first hold on retains current figure error inside try/catch block when adding new stuff dbclear clears breakpoints hold off restores to default dbcont resume execution (no hold on) lasterr Last error message set(h,’WindowStyle’,’Docked’); Docked window lastwarn Last warning message style for plots break Terminates executiion of for/while loop datetick(’x’,yy) time series axis waitbar Waiting bar plotyy(x1,y1,x2,y2) plot on two y axis refreshdata refresh data in graph if specified source Data import/export drawnow do all in event queue xlsread/xlswrite Spreadsheets (.xls,.xlsm) * Some markers: ’, +, *, x, o, square readtable/writetable Spreadsheets (.xls,.xlsm) * Some colors: red, blue, green, yellow, black dlmread/dlmwrite text files (txt,csv) * color shortcuts: r, b, g, y, k load/save -ascii text files (txt,csv) * Some line styles: -, --, :, -. load/save matlab files (.m) * shortcut combination example: plot(x,y,’b--o’) imread/imwrite Image files fig1 = plot(x,y) set(fig1, ’LineWidth’, 2) set(fig1, ’LineStyle’, ’-’) set(fig1, ’Marker’, ’.’) set(fig1, ’color’, ’red’) set(fig1, ’MarkerSize’, 10) set(fig1, ’FontSize’, 14) figure figure(j) get(j) Output commands format short format long disp(x) disp(x) num2str(x) num2str([’nA is = ’ num2str(a)]) mat2str(x) int2str(x) sprintf(x) Displays 4 digits after 0 Displays 15 digits after 0 Displays the string x Displays the string x Converts the number in x to string OFTEN USED! ! Converts the matrix in x to string Converts the integer in x to string formated data to a string System commands addpath(string) genpath(string) pwd mkdir tempdir inmem exit dir ver adds path to workspace gets strings for subfolders Current directory Makes new directory Temporary directory Functions in memory Close matlab list folder content lists toolboxes Programming commands return Return to invoking function exist(x) checks if x exists G=gpuArray(x) Convert varibles to GPU array function [y1,...,yN] = myfun(x1,...,xM) Anonymous functions not stored in main programme myfun = @(x1,x2) x1+x2; or even using myfun2 = @myfun(x) myfun(x3,2) Conditionals and loops for i=1:n procedure end while(criteria) procedure end Iterates over procedure incrementing i from 1 to n by 1 Iterates over procedure as long as criteria is true(1) if(criteria 1) procedure1 elseif(criteria 2) procedure2 else procedure3 end switch switch_expression case 1 procedure 1 case 2 procedure 2 otherwise procedure 3 end if criteria 1 is true do procedure 1 ,else if criteria 2 is true do procedure 2 , else do procedure 3 if case n holds, run procedure n. If none holds run procedure 3 (if specified) General comments • Monte-Carlo: If sample sizes are increasing generate largest size first in a vector and use increasingly larger portions for calculations. Saves time+memory. • Trick: Program that (1) takes a long time to run and (2) doesnt use all of the CPU/memory ? - split it into more programs and run using different workers (instances). • Matlab is a column vector based language, load memory columnwise first always. For faster code also prealocate memory for variables, Matlab requires contiguous memory usage!. Matlab uses copy-on-write, so passing pointers (adresses) to a function will not speed it up. Change variable class to potentially save memory (Ram) using: int8, int16, int32, int64, double, char, logical, single • You can turn the standard (mostly) Just-In-Time compilation off using: feature accel off. You can use compiled (c,c++,fortran) functions using MEX functions. • Avoid global variables, they user-error prone and compilers cant optimize them well. • Functions defined in a .m file is only available there. Preface function names with initials to avoid clashes, eg. MrP function1. • Graphic cards(GPU)’s have many (small) cores. If (1) program is computationally intensive (not spending much time transfering data) and (2) massively parallel, so computations can be independent. Consider using the GPU! • Using multiple cores (parallel computing) is often easy to implement, just use parfor instead of for loops. • Warnings: empty matrices are NOT overwritten ([] + 1 = []). Rows/columns are added without warning if you write in a nonexistent row/column. Good practise: Use 3i rather than 3*i for imaginary number calculations, because i might have been overwritten by earlier. 1/0 returns inf, not NaN. Dont use == for comparing doubles, they are floating point precision for example: 0.01 == (1 − 0.99) = 0. Copyright  c 2015 Thor Nielsen (thorpn86@gmail.com) http://www.econ.ku.dk/pajhede/