Introduction to mdadata class
Class mdadata extends usability of conventional matrices, allowing to keep row and column names with data values and use them when show or plot the data. The mdadata can be used only with numerical values.
Every object of mdadata class has following properties, which can be set by a user when creating:
'values' — data values, a matrix with numbers
`rowNames` a cell array with row names (optional)
`colNames` a cell array with column names (optional)
`dimNames` a cell array with names for each of the two dimensions (optional)
`name` a text string with short name of the dataset (optional)
The default syntax for creating the dataset object is:
data = mdadata(values, objNames, colNames, dimNames, name);
Most of the properties are optional, they will be generated automatically or remain empty if user does not provide the proper values:
% Create a dataset with default property values:
d = mdadata([180 85; 172 68; 156 50]);
show(d)
Variables
1 2
---- ---
180 85
172 68
156 50
disp('Dim names are:')
disp(d.dimNames)
Dim names are:
'Objects' 'Variables'
Function show() shows dataset values as a table with column and row names. As one can see column names were generated as '1', '2' and so on and row names remain empty. Default dimnames values are 'Objects' and 'Variables' and dataset name is not defined. The row names and column names must be unique!
Here is an example with all options.
% create a dataset with all properties provided
d = mdadata([180 85; 172 68; 156 50], {'Lars', 'Peter', 'Lena'}, ...
{'Height', 'Weight'}, {'People', 'Parameters'}, 'People data');
show(d)
People data:
Parameters
Height Weight
------- -------
Lars 180 85
Peter 172 68
Lena 156 50
Column and row names should consist only of Latin letters and/or numbers. However you can provide them in a free form (e.g. ('Height of person, cm') and the name will be converted as following: spaces will be removed and every word capitalized, if there are punctuation marks or parentheses they will be also removed as well as the part of the name after the first punctuation symbol (for the example above it will become 'HeightOfPerson').
This is needed to identify the names correctly when subset the datasets with logical expressions. The original (user provided) names will be also kept and used as labels when you print or plot the data.
d = mdadata([180 81; 172 66; 156 48], {'Lars Larsen', 'Peter from room 22', 'Lena'}, ...
{'Height', 'Body mass (kg)'});
show(d)
Variables
Height Body mass (kg)
------- ---------------
Lars Larsen 180 81
Peter from room 22 172 66
Lena 156 48
disp('Column names are:')
disp(d.colNames)
Column names are:
'Height' 'Bodymass'
disp('Row names are:')
disp(d.rowNames)
Row names are:
'LarsLarsen' 'Peterfromroom22' 'Lena'
If it is needed to specify only selected parameters, the others should be provided as empty arrays:
% create a dataset with variable and dimension names provided
d = mdadata([180 85; 172 68; 156 50], [], {'Height', 'Weight'}, {'People', 'Parameters'});
show(d)
Parameters
Height Weight
------- -------
180 85
172 68
156 50
Names and values can be changed for existent object:
d = mdadata([180 85; 172 68; 156 50]);
d(:, 1).colNames = {'Height'};
d.dimNames = {'People', 'Parameters'};
d(1, :).values = [181 88];
show(d)
`
Parameters
Height 2
------- ---
181 88
172 68
156 50
Rows and columns can be removed using specific methods.
d = mdadata([180 85 20; 172 68 20; 156 50 21; 161 51 22]);
show(d)
Variables
1 2 3
---- --- ---
180 85 20
172 68 20
156 50 21
161 51 22
d.removecols(1);
show(d)
Variables
2 3
--- ---
85 20
68 20
50 21
51 22
d.removerows(2:3);
show(d)
Variables
2 3
--- ---
85 20
51 22
You already know about function show(), it has an additional parameter, which defines how many significant digits to show (default value is 3):
d = mdadata([180 0.02345; 175 0.00123]);
show(d)
Variables
1 2
---- --------
180 0.0234
175 0.00123
show(d, 2)
Variables
1 2
-------- -------
1.8e+02 0.023
1.8e+02 0.0012
show(d, 5)
Variables
1 2
---- --------
180 0.02345
175 0.00123