How to extract digits

Hello everybody. I would like to know how to extract the first digit of each element of a vector.

Example

a = [438;259;21;4;79876]

out = [4;2;2;4;7]

At the same time I would like to know if there is a way to extract the first two digits of each element:

a = [438;259;21;44;79876]

out = [43;25;21;44;79]

tnxxx

Answer 1

[Voss]

Assuming they're all positive numbers, you can try this (the 2-digit result won't be accurate for one-digit numbers):

a = [438;259;21;44;79876];

power10 = 10.^floor(log10(a));
first_digit = floor(a./power10)

first_digit = 5×1

4 2 2 4 7

first_two_digits = floor(a./power10*10)

first_two_digits = 5×1

43 25 21 44 79

[LC]

Hi, very thanks for your help! I also wanted to ask you another question: if instead of a vector with numbers I had some text strings, how could I extract the first element?

Ex

a=['ab';'cde','hhh','kbj']

Error using vertcat
Dimensions of arrays being concatenated are not consistent.

out=['a';'c';'h';'k']

tnxxxx

[Image Analyst]

LC, that syntax is not right. What do you really have? A vector of strings (newer and less common)? Or a cell array of character arrays (older and more common)? They are different.

[LC]

Hi, sorry if I've been unclear, I'm just starting out with matlab. I believe it is an array of characters. Now I'm sending you a photo of my variable so maybe it's clearer. Thanks.

[Voss]

That is indeed an array of characters (in particular, it is a character vector - a vector being a 1-dimensional array; in this case it is a row vector, i.e., of size 1-by-something).

Here's how you could extract the first digit (or first two digits) of each number in there, resulting in a cell array of character vectors or a character array.

(The idea is to read the character vector ST using sscanf() to get numbers and then use the approach in my answer to get the numeric digit(s), then 'print' those digit(s) to into a cell array or character array using sprintfc() or sprintf().)

ST = '51         51         50        49         49';

a = abs(sscanf(ST,'%f'))

a = 5×1

51 51 50 49 49

power10 = 10.^floor(log10(a));
first_digit = sprintfc('%d',floor(a./power10))

first_digit = 5×1 cell array
    {'5'}
    {'5'}
    {'5'}
    {'4'}
    {'4'}

first_two_digits = sprintfc('%d',floor(a./power10*10))

first_two_digits = 5×1 cell array
    {'51'}
    {'51'}
    {'50'}
    {'49'}
    {'49'}

Those are the cell array results, and here's the character array version (with one or two columns depending on if you're taking one or two digits):

first_digit = sprintf('%d',floor(a./power10)).'

first_digit = 5×1 char array
    '5'
    '5'
    '5'
    '4'
    '4'

first_two_digits = reshape(sprintf('%d',floor(a./power10*10)),2,[]).'

first_two_digits = 5×2 char array
    '51'
    '51'
    '50'
    '49'
    '49'

Now, let's say ST was like this instead:

ST = '-5.1         +5.1e+002         50.0        -49e-3         49';

Then the approach above will still work:

a = abs(sscanf(ST,'%f'))

a = 5×1

5.1000 510.0000 50.0000 0.0490 49.0000

power10 = 10.^floor(log10(a));
first_digit = sprintfc('%d',floor(a./power10))

first_digit = 5×1 cell array
    {'5'}
    {'5'}
    {'5'}
    {'4'}
    {'4'}

first_two_digits = sprintfc('%d',floor(a./power10*10))

first_two_digits = 5×1 cell array
    {'51'}
    {'51'}
    {'50'}
    {'49'}
    {'49'}

first_digit = sprintf('%d',floor(a./power10)).'

first_digit = 5×1 char array
    '5'
    '5'
    '5'
    '4'
    '4'

first_two_digits = reshape(sprintf('%d',floor(a./power10*10)),2,[]).'

first_two_digits = 5×2 char array
    '51'
    '51'
    '50'
    '49'
    '49'

But let me reiterate that this approach still has the limitation that the two-digit result for a one-digit number will not be accurate ('4' will return '40' for instance).

In general, the best choice of method will depend on:

• what format(s) the numbers in the ST character vector might possibly be in, and
• what class of variable you want to end up with, e.g., cell array of character vectors, character array, string array, numeric array

[Stephen23]

ST = '51         51         50        49         49';
DD = regexp(ST,'\d\d','match')

DD = 1×5 cell array
    {'51'}    {'51'}    {'50'}    {'49'}    {'49'}

[LC]

thank you very much for your support, it works great because I never have a one-digit number.