本文主要包括以下内容
- 数据类型
- 文件输入输出
数据类型
7.1 Introduction to Data Types
The Limitation of Computers Real numbers in mathematics: Can be infinitely large Have infinitely fine resolution Computers: Finite memory Upper limit on the largest number that can be represented Lower limit on the absolute value of any non-zero number The set of values that can be represented by a MATLAB variable is finite.
Data Types
MATLAB: many different data types
A data type is defined by:
Set of values
Set of operations that can be performed on those values
MATLAB:
All elements of a given array must be of the same type
Elementary type
Type of a MATLAB array is defined by
Number of dimensions
Size in each dimension
Elementary type
to check data type: class(x)
x = 23; class(x) = double — default data type for all numerical data in matlab (incl. complex numbers)
whos: check type & size of every variables
class(double) = char — character
class(3<4) = logical
Numerical Types
double
Default type in MATLAB
Floating point representation
Example: 12.34 = 1234 * 10-2
Mantissa and exponent
64 bits (8 bytes)
single
32-bit floating point
Integer types
Signed (both negative & non-negative numbers), unsigned (only non-negative numbers)
8-, 16-, 32-, 64-bit long
Range of Values
Inf: “Infinity” — x/0
NaN: “Not-a-Number” — 0/0
Useful functions
Type check:
class
isa
>> isa(x,’double’)
Range check:
intmax, intmin
realmax, realmin
>> intmax(’uint32’)
Conversion:
Name of function = name of desired data type
int8(x), uint32(x), double(x), etc.
Operators
Arithmetic operators
Operands of the same data type: No problem
Different data types:
“mixed-mode arithmetic”
Many rules and restrictions
Relational operators
Different data types are always allowed
Result is always of logical type
Conversion In MATLAB
>> k = uint8(500)
k =
255
>> k = uint8(256)
k =
255
>> k = uint8(-1)
k =
0
若赋值超出变量范围，则会自动变为最接近变量范围的值(clipping)
Range Check in MATLAB
>> intmax('uint8')
ans =
255
>> intmin('uint8')
ans =
0
>> intmin('int32')
ans =
-2147483648
>> realmax('double')
ans =
1.7977e+308
>> realmin('single')
ans =
1.1755e-38
7.2 Strings
In MATLAB: char(x)输出ASCII里第x位的符号（e.g. function char_codes） Strings和数集一样可以用length
>> MOOC_title = 'MATLAB for Smarties’;
>> length(MOOC_title)
ans =
19
>> MOOC_title(1)
ans =
M
String functions
findstr:
>> MOOC_title = 'MATLAB for Smarties';
>> strfind(MOOC_title,'LAB')
ans =
4
输出的是第一个字母的位置
case sensitive
>> strfind(MOOC_title,'r')
ans =
10 15
多次出现的话会输出多个值
strcmp:
>> MOOC_title = 'MATLAB for Smarties';
>> lang = ‘MATLAB';
>> strcmp(MOOC_title, lang)
ans =
0
Meaning two strings are not the same.
case sensitive
>> strcmp(MOOC_title(1:6), lang)
ans =
1
strcmpi: to compare while ignore case
>> strcmpi(MOOC_title(1:6), 'Matlab')
ans =
1
str2num (num2str):
>> pi_number = 3.1416
pi_number =
3.1416
>> pi_digits = '3.1416'
pi_digits =
3.1416
>> str2num(pi_digits)
ans =
3.1416
sprintf:
与fprintf一样，除了有一个输出值可以赋予变量（变量类型为string/char)
7.3 Structs
Structs
An array must be homogeneous:
It cannot contain elements of multiple types.
A struct can be heterogeneous:
It can contain multiple types.
A struct is different from an array: fields, not elements field names, not indices Fields in the same struct can have different types. Versatility inside: A field of a struct can contain another struct. Structs can hold arrays, and arrays can hold structs. Struct functions
Create a struct
>> account.number = 1234567
account =
number: 1234567
>> account.balance = 5000
account =
number: 1234567
balance: 5000
>> account.owner.name = 'Joe Smith'
account =
number: 1234567
balance: 5000
owner: [1x1 struct]
>> account.owner.email = 'joe@matlabcooc.com'
account =
number: 1234567
balance: 5000
owner: [1x1 struct]
Create another struct with the same name — 和原先的有一样的field，原先的自动变成（1）；只限于一级子类别，对account.owner.后面的类别没影响
>> account(2).number = 7654321
account =
1x2 struct array with fields:
number
balance
owner
isfield: 输出真或假（有或无）
>> account(1:2).owner
ans =
name: 'Joe Smith'
email: 'joe@matlabcooc.com'
ans =
name: 'Jane Farmer'
age: 23
>> isfield(account(2).owner,'age')
ans =
1
>> isfield(account(1).owner,'age')
ans =
0
rmfield: remove field
>> account(1).owner = rmfield(account(2).owner,'age’);
>> account(1).owner
ans =
name: 'Jane Farmer'
struct:
>> course = struct('Area', 'CS', 'number', 103, 'title', 'Introductory Programming for Engineers and Scientists')
course =
Area: 'CS'
number: 103
title: 'Introductory Programming for Engineers and Scientists'
7.4 Cells
Pointers
How to store a page of text?
Each line should be a separate string
Cannot use an array of chars:
Each line would have to have the same length
A vector of objects with each referring to one line
Pointer
Each variable (scalar, vector, array, etc.) is stored in the computer memory.
Each memory location has a unique address.
A pointer is a variable that stores an address.
MATLAB calls a pointer a “cell”.
Cells
MATLAB has a restrictive pointer model
Strict rules on what can be done with cells
Harder to make mistakes
But it is a powerful way to store heterogeneous data
Cell arrays
Used more frequently than structs
New syntax:
To access the data a cell points to, use: { }
若令c2 = c1，改变c1中的元素，c2的元素不会同步变化——c2中的元素只是c1元素的copy
Cell functions,作为参数传入函数中，发生改变，原cell指向的不变。
文件输入输出
8.1 File Input / Output
File Input / Output
File:
Area in permanent storage (disk drive)
Stores information
Managed by the operating system
Can be copied or moved
Can be accessed by programs
File Input/Output (I/O)
Data exchange between programs and computers
Data exchange between the physical world and computers
Saving your work so you can continue with it later
MATLAB can handle
Mat-files and M-files AND text, binary, and Excel files
>> pwd: print the directory of current folder
>> ls: print all files in the current folder
>> cd(x): change current folder - 可以是完整的路径，也可以输入子文件夹的名字，也可以用cd(‘..’)回到母文件夹，cd(‘../..’)回到再上一级文件夹
>> mkdir('new_folder’): make a new folder
>> rmdir('new_folder’): remove a new folder
>> save: 保存工作区内容到matlab.mat
>> load: 载入matlab.mat
>> save my_data_file data s a: 可以自己设定储存的名字（my_data_file)和保存的变量（data s a)
8.2 Excel Files
Excel files
Microsoft Excel® is a widely used data-analysis tool
Many other programs support reading and writing Excel files
MATLAB does too with two built-in functions
xlsread
xlswrite
Reading Excel files
>> [num,txt,raw] = xlsread('Nashville_climate.xlsx');
num: smaller than the size of spreadsheet, incl. NaN (Not a Number)
>> [~, text] = xlsread('Nashville_climate_data.xlsx’): ignores the numerical data, only keep the text data
>> [~,~, everything] = xlsread('Nashville_climate_data.xlsx’): ignores the numerical & text, only show the whole file
>> num = xlsread('Nashville_climate_data.xlsx',1,'D15’): 得到表中的第一个sheet里D15的内容
‘D15’可以换成’D15:E17’用来输出一个区域的值（左上角/右下角）
8.3 Text Files
Text files
Text files contain characters
They use an encoding scheme:
ASCII or ◦ Any one of many other schemes
MATLAB takes care of encoding and decoding
Before using a text file, we need to open it
Once done with the file, we need to close it
Opening text files
Opening: fid = fopen(filename, permission) — fid = file identifier
Closing: fclose(fid)
fid: Unique file identifier for accessing file
Permission: what we want to do with the file—
read, write, overwrite, append, etc.
例子： write_temp_precip_txt
Reading text files
One line at a time
type prints a text file in the command window
Let’s re-implement it:
function view_text_file(filename)
fid = fopen(filename,'rt');
if fid < 0
error('error opening file %s\n', filename);
end
% Read file as a set of strings, one string per line:
oneline = fgets(fid);
while ischar(oneline)
fprintf('%s',oneline) % display one line
oneline = fgets(fid);
end
fprintf('\n');
fclose(fid);
Reading lines into string variables is easy
Parsing these strings to get numerical data is much harder
Not covered
Binary files are more suited for numerical data
8.4 Binary Files
Binary files
Binary file = “not a text file”
Many different ways to represent numbers
All we need to know are their types.
Binary files need to be Opened with fopen Closed with fclose Writing binary files Data type is important
Example: write a double array into a binary file
function write_array_bin(A,filename)
fid = fopen(filename,'w+');
if fid < 0
error('error opening file %s\n', filename);
end
fwrite(fid,A,'double');
fclose(fid);
Reading binary files
Example: read a double array from a binary file
function A = read_bin_file(filename,data_type)
fid = fopen(filename,'r');
if fid < 0
error('error opening file %s\n',filename);
end
A = fread(fid,inf,data_type);
fclose(fid);
Inf - open all the file
如果纯粹的用fwrite和fread，会使format改变，变成一个向量而无矩阵等，这种时候就要将格式信息也加进去—— write_dims_array_bin(A,filename) & A = read_dims_array_bin(filename)