MatLab Cheatsheet

table.Properties              Access table properties
t{i}                          Access by index
t(2:end,[1, 5])               From t make new table w/ specific cols/rows
        {'var1','var2'}       Also valid form to access cols
t.newVar =                    Add new col to table
t = datetime(timestamp)       Create datetime table
  hour(t), day(t), ...
t{[i j],'colName'} = {'';''}  Modify values in position i,j of column in table

when sorting a list that has a relationship with another list, 
one must also obtain the indexed values of the sort so that the relationship between both lists is kept:
x,y;
[sorted_x, idx_vals] = sort(x);
sorted_y = y(idx_vals);
plot(sorted_x,sorted_y);

enums -> categorical(list). more memory efficient
         categories(list) make array of categories found in list
         mergecats(list,{'','',''},biggerCategory)

useful flags:
Omitnan
reading tables: HeaderLines, CommentStyle

setoperations:
        setdiff(a,b)    return everything in a but not in b
        

useful functions:
nnz:            count non zero matrix elements
any:            determine which row/cols have a true value
ismissing:      find all NaN, NaT, undefined indexes in a matrix
isnan:          same as ismissing but only for NaN

When importing the data:
        The numeric missing values are replaced by NaN. 
        The missing datetime values are replaced by NaT (not a time)
        The missing categorical values are replaced by undefined.
        
Ex: 
        valid_cats = categories(data.Var);
        valid_cats = setdiff(valid_cats, 'N/A');
        data.NewVar = mergecats(data.Var, valid_cats, 'Land');
        data.NewVar = renamecats(data.Var, 'N/A', 'Sea')
        
 Discretize: discretize(data.Var,bins,'Categorical',catnames)
 
function handle = reference to a function. Allows to pass fct as argument -> fct = @fct_name

When creating new table () extracts as table -> variable info but can't perform operations
                        {} extracts as cell  -> can perform operations but no variable info

Main Datatypes:
Arrays: Use when heterogenous datatypes and no need for math ops.
  - Struct: data accessible by name.                        s = struct()
  - Cell:   data accessible by index.                       c = {}
            diff. sized strings  s = {'', '', ''}
            add/del/mod value -> c{i} = ''
Matrix: Use when homogenous datatypes and lots of math ops. v = []
            same sized strings s = ['', '', '']
            add/del cell -> c(i) = []

Table:  Use when heterogenous variables but same col size.  t = table()
Readtable to import from spreadsheet
Joining Tables: 
  [T1;T2]             Concatenation 
  join(T1,T2) 	      The key values in the first input all exist in the second input. If not, then reverse the input order.
  innerjoin(T1,T2)    inner join -> (AND)
  outterjoin(T1,T2'Mergekeys',true)   Outter join -> (OR). All empty values NaN
  
  Basic graph properties:
        hold	        Plot replacement behavior of axes
        xlim/ylim/zlim	Limits of the appropriate axis
        grid	        Axes grid lines
        axis	        Axis limits, shape, and appearance
        colormap	Map used to assign indexed colors in a figure
        view	        3-D viewpoint of axes
        axis() square, tight, auto, fill 
        
 DATASTORE:
 
 when trying to import data into a table from multiple files
 1. create data store
 2. adjust data store properties
 3. import data
        
 GRO_UPING DATA BY VARIABLES
[grp_nums,grp_vals] = findgroups(v) -> given list, map to 1,2,3.... 
histcounts(list) -> histogram of number of occurrences of group in list

splitapply(@fct,data,groups): split list by afor created groups and apply input function

MANIPULATING PLOTS
fig = gcf
ax = gca
obj = gco
 
========================================================================
% Read Table
data = readtable('fuelEconomy.txt','HeaderLines', 4);

% Find and delete NaN rows
nan_idx = ismissing(data.CombinedMPG);
data(nan_idx,:) = [];

% Categorize variable
bins = [0 20 30 70];
cat_names = {'Low','Medium','High'};
MPGClass = discretize(data.CombinedMPG,bins,cat_names);
MPGClass = categorical(MPGClass);

low_idx = MPGClass == 'Low';
medium_idx = MPGClass == 'Medium';
high_idx = MPGClass == 'High';

% Plot
scatter(data.CityMPG(low_idx), data.HighwayMPG(low_idx),8,'r','filled');
hold on;
scatter(data.CityMPG(medium_idx), data.HighwayMPG(medium_idx),8,'b','filled');
scatter(data.CityMPG(high_idx), data.HighwayMPG(high_idx),8,'k','filled');
legend('Low','Medium','High');
grid('on');
xlabel('City MPG');
ylabel('Highway MPG');
hold off;


##### FANCY PLOT

% Read data
dat = datastore('fuelEconomy2.txt');
dat.ReadSize = 362;
data = dat.read;

% Group by number of cylinders
[gNum,gVal] = findgroups(data.NumCyl);

% Find average by groups
avgMPG = splitapply(@mean,data.CombinedMPG,gNum);

% Create a bar chart
b = bar(avgMPG);
xlabel('Number of cylinders')
title('Average MPG')

% Customize the chart
f = gcf;
a = gca;

f.Color = [0.81 0.87 0.9];

a.Color = [0.81 0.87 0.9];
a.Box = 'off';
a.YAxisLocation = 'right';
a.YGrid = 'on';
a.GridColor = [1 1 1];
a.GridAlpha = 1;
a.XTickLabel = gVal;
a.YLim = [0 40];

ax = a.XAxis;
ax.TickDirection = 'out';

b.FaceColor = [0,0.31,0.42];
b.BarWidth = 0.5;

##### SURF

%% Create an XY grid from the raw data
xv = min(x):0.01:max(x);
yv = min(y):0.01:max(y);
[X,Y] = meshgrid(xv,yv);

%% Interpolate the raw data and evaluate it on the X,Y grid
Z = griddata(x,y,z,X,Y);

##### PCOLOR
The pcolor function has the same syntax as surf. In fact, it actually creates a flat surface with ZData all set to 0 and CData set to Z.
>> pcolor(X,Y,Z)


### LOOPING THROUGH FILE
%% Open file
fid = fopen('hurricaneData1960.txt');

%% TODO: Create a loop that reads the data until the end of the file
while ~feof(fid)
    % Read header line
    headerLine = fgetl(fid);
    % Split up the line
    txtCell = strsplit(headerLine);
    % Find locations of M and =
    mWhere = strcmp(txtCell,'M');
    eqWhere = strcmp(txtCell,'=');
    % Find the index value that follows 'M ='
    idx = all([0 0 mWhere(1:end-2); 0 eqWhere(1:end-1)]);
    % Assign that number to nLines
    nLines = txtCell{idx};
    nLines = str2double(nLines);

    % Skip the next line
    fgetl(fid);

    % Read in nLines of data
    data = textscan(fid,'%D%f%f%f%f',nLines,'Delimiter','\t');
    % Read the rest of the line
    fgetl(fid);
end

%% All done. Close the file
fclose(fid);