Switch case on an array

57 次查看（过去 30 天）
luca
luca 2019-9-19
编辑： Guillaume 2019-9-19
Given the following code
TA = 0
TB = 0
TC = 0
x = [12 64 24];
s = num2str(x);
n = length(s)
for k=1:n
switch s(k)
case 12
TA = TA + 1;
case 24
TB = TB + 1;
case 64
TC = TC + 1;
end
end
I would like to create a switch case depending on the value on x.
Following the code, the final result should be
TA = 1
TB = 1
TC = 1
But I don't know how to write x in the right character array, and the code give me a wrong result.
Which is the most clever way to solve this problem?
  4 个评论
显示 2更早的评论
Guillaume
Guillaume 2019-9-19
<pedantic> s is a char array not an array of strings which is a different thing since R2016b. </pedantic>
It's difficult to answer the question without an explanation of why s is char array instead of the more useful numeric vector x. We need an explanation of what the whole purpose of the code is.
As it is the code presented looks very badly designed, with unnecessary number to string conversions and sequentially named variables.
luca
luca 2019-9-19
编辑：luca 2019-9-19
The code has the following purpose:
I have a vector
x = [12 24 64 24 12 12]
and three quantity intially equal to 0
TA = 0;
TB = 0;
TC = 0;
I want to iterate on x's columns. Everytime I find 12 I want to increment of 1 TA, everytime i find 24 I want to increment by 1 TB and everytime I find 64 I want to increment by 1 TC.
In this case we should obtain
TA = 3
TB = 2
TC = 1
I hope I have been clearer now

请先登录，再进行评论。

请先登录，再回答此问题。

采纳的回答

Guillaume
Guillaume 2019-9-19
Yes, it's a lot clearer. So, the transition to a char array is a complete red herring. It's totally unnecessary.
Using sequentially named variables is a bad idea. It will force you to write complicated code. If you have 10 different numbers that you want to count would you add another 7 cases to your switch? How about 100 numbers, 10000 numbers. Will you write all these cases?
The answer is to use array indexing, T(1) the count of your first, T(2) the count of your 2nd number, etc. Then you can use a loop. For example,one (inneficient) way to then implement your code would be:
%input
x = [12 24 64 24 12 12]
lookup = [12, 24, 64]; %values to compute the histogram of
%code
count = zeros(size(lookup))
for i = 1:numel(x)
[found, where] = ismember(x(i), lookup);
if found
count(where) = count(where) + 1;
end
end
But the whole thing can be achieve without loops, if or case. You want to compute an histogram, use the histcounts function:
x = [12 24 64 12 12];
values = unique(x)'; %get unique values of
count = histcounts(x, [values; Inf])'; %inf is right edge of last bin
table(values, count) %for pretty display
returns
ans =
3×2 table
values count
______ _____
12 3
24 1
64 1
  2 个评论
luca
luca 2019-9-19
and then I can write
T1= count(1)
T2= count(2)
T3=count(3)
to obtain the value that I want. right?
Thanks Guillaume for all the explanation and all the details.
Guillaume
Guillaume 2019-9-19
编辑：Guillaume 2019-9-19
You can but you shouldn't. Again, what if you want 1000 values, you'll be writing Txx = count(xx) a 1000 times.
Simply, continue using count(1) wherever you were going to use T1. Don't number variables, use indexing.

请先登录，再进行评论。

更多回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Characters and Strings 的更多信息

标签

产品


版本

R2019b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by