How do you get up to first n characters in a string?

I have an array of strings of variable length s_arr. I want to limit the strings to only contain the first 3 elements. Some of the string have less than 3 elements, in which case they should be unchanged. I tried

arrayfun(@(s) s(1:3), s_arr)

but this gives an error when it finds a string with fewer than 3 elements. Why does MATLAB give an error and not just return the string unchanged? How can I achieve this goal?

[Stephen23]

You were very close:

s_arr = ["a";"ab";"abc";"abcd";"abcde"]

s_arr = 5×1 string array
    "a"
    "ab"
    "abc"
    "abcd"
    "abcde"

cellfun(@(s) s(1:min(end,3)), s_arr, 'uni',0)

ans = 5×1 cell array
    {'a'  }
    {'ab' }
    {'abc'}
    {'abc'}
    {'abc'}

Answer 1

[Dyuman Joshi]

A simple approach using extractBetween -

s_arr = ["a";"ab";"abc";"abcd";"abcde";"abcdef";"xyzxyzxyz"]

s_arr = 7×1 string array
    "a"
    "ab"
    "abc"
    "abcd"
    "abcde"
    "abcdef"
    "xyzxyzxyz"

pos = 4; % pos  = no. of characters + 1

extractBetween(s_arr, 1, min(pos, strlength(s_arr)))

ans = 7×1 string array
    "a"
    "ab"
    "abc"
    "abcd"
    "abcd"
    "abcd"
    "xyzx"

Answer 2

[Paul]

If a string array:

s_arr = ["a";"ab";"abc";"abcd";"abcde"]

s_arr = 5×1 string array
    "a"
    "ab"
    "abc"
    "abcd"
    "abcde"

s_arr(strlength(s_arr)>3) = extractBefore(s_arr(strlength(s_arr)>3),4)

s_arr = 5×1 string array
    "a"
    "ab"
    "abc"
    "abc"
    "abc"

Exact same code works on a cell array

s_arr = cellstr(["a";"ab";"abc";"abcd";"abcde"])

s_arr = 5×1 cell array
    {'a'    }
    {'ab'   }
    {'abc'  }
    {'abcd' }
    {'abcde'}

s_arr(strlength(s_arr)>3) = extractBefore(s_arr(strlength(s_arr)>3),4)

s_arr = 5×1 cell array
    {'a'  }
    {'ab' }
    {'abc'}
    {'abc'}
    {'abc'}

[Stephen23]

+1 This gracefully handles the shorter elements. Stylistically I would define the index beforehand:

s_arr = ["a";"ab";"abc";"abcd";"abcde"]

s_arr = 5×1 string array
    "a"
    "ab"
    "abc"
    "abcd"
    "abcde"

N = 3;
X = strlength(s_arr)>N;
s_arr(X) = extractBefore(s_arr(X),1+N)

s_arr = 5×1 string array
    "a"
    "ab"
    "abc"
    "abc"
    "abc"

Answer 3

[Image Analyst]

% Character array
charArray = 'abcdefgh'	% Single quotes
n = 3
out = charArray(1:n) % Another way
% String
str = "abcdefgh" % Double quotes
n = 3
charArray = char(str)
out = charArray(1:n)

charArray =

'abcdefgh'

n =

3

out =

'abc'

str =

"abcdefgh"

n =

3

charArray =

'abcdefgh'

out =

'abc'

[Image Analyst]

For the case where n is longer than the string, or less than the string (so more general and robust):

% Character array
charArray = 'ab'	% Single quotes
n = 30
out = charArray(1:min(length(charArray), n)) % Another way
% String
str = "ab" % Double quotes
charArray = char(str)
out = charArray(1:min(length(charArray), n))

Answer 4

[Ramanan]

You can use the extractBefore function in the array function instead.

pos = 4; % pos  = no. of characters + 1
out = arrayfun(@(s) extractBefore(s,pos), s_arr);

[Stephen23]

This fails due to the shorter elements:

s_arr = ["a";"ab";"abc";"abcd";"abcde"]

s_arr = 5×1 string array
    "a"
    "ab"
    "abc"
    "abcd"
    "abcde"

pos = 4; % pos  = no. of characters + 1
out = arrayfun(@(s) extractBefore(s,pos), s_arr)

Error using extractBefore
Numeric value exceeds the number of characters in element 1.

Error in solution>@(s)extractBefore(s,pos) (line 3)
out = arrayfun(@(s) extractBefore(s,pos), s_arr)

See Paul's answer from three years ago for one approach to handle this.

[Ramanan]

You can add the min function and make use of the length of the string inside the arrayfun.

I don't notice the previously answered one.

s_arr = ["a";"ab";"abc";"abcd";"abcde";"abcdef";"xyzxyzxyz"];
pos = 4; % pos  = no. of characters + 1
out = arrayfun(@(s) extractBefore(s,min(pos,strlength(s))), s_arr)

out = 7×1 string array
    ""
    "a"
    "ab"
    "abc"
    "abc"
    "abc"
    "xyz"

Also I think a more simplified version is submitted by Dyuman Joshi.

Answer 5

[Geoff]

Learning about and using regex (regular expressions) has solved many string dilemmas for me. Here's a regex one liner for this question:

regexprep(s_arr,"(.{3}).*","$1")

Explanation:

The function regexprep ("reg ex replace") is just a fancy text search-and-replace. In a nutshell, our strategy is to find the entire string and replace it with just its first {3} characters.

The first input s_arr is your string array. regexprep will operate on each element of the array.

The next two inputs are our search "(.{3}).*" and our replacement "$1". You can read about regex syntax here. In a nutshell:

• The period . matches any character.
• And .{3} matches any 3 consecutive characters.
• And .* matches any number of characters -- i.e. the rest of the string.

So, we've searched for the entire string, represented it in two parts: the first 3 characters (.{3}) and the rest .*. The surrounding parentheses in (.{3}) say we want to capture (i.e. remember) those characters. The $1 in our replacement string is what recalls those captured characters.

So, the entire string has been replaced by the first three characters. If the string is less than 3 characters, regexprep doesn't find any matches in its search, so nothing is replaced.

Pros and cons:

Con: it's cryptic.

Pro: It doesn't require Matlab to compute the length of the strings in the array.