Validate and get exact number of characters using the regular expression

Hi all,

I am using MATLAB R2016a version and working on a chacters set like

str = 'ENGINE-45'; % this is okay(digits are 2 after hyphen)

where i need to validate and get as numbers of digits after hyphen (-), it should be 2 digtits but in some case it is 3, 4, 5 number digits like

str = 'ACCCAT-455'; % this is not okay(3 digits)
str = 'VCCASNSRR-12344'; % this is not okay(5 digits)

I have used reguler expresion to get charcters and digits as

out_str = regexp(str, '[A-Z]+-[0-9]{2}', 'match');
ans = 

    'VCCASNSRR-12'  % used {2} but where inputs are 5, tryiing to restrict this to only 2 char

but this is not what i required i need to get if and only 2 digits after hyphen?

[dpb]

Two exactly, not <=2?

'VCCASNSRR-1'  % OK or not?

Answer 1

[J. Alex Lee]

Try

out_str = regexp(str, '[A-Z]+-[0-9]{2}$', 'match');

[J. Alex Lee]

For completeness, you can still use regex for the "numbers only" problem using look-behind and making only minimal changes to original expression. Surely it's not the only way, I'm also pretty lame at regex. I'm sure only aliens can actually make regex expressions on demand without consulting docs :) Also, now I know why I had trouble with word boundaries, I didn't know about \> and \<

candidates = [...
	"SOMETHING-2";...
	"ENGINE-45";...
	"ACCCAT-455";...
	"ANOTHER-23";...
	"PREFIX BLAH-64";...
	"BLAH-75 SUFFIX";...
	"VCCASNSRR-12344"]

FullMatch = regexp(candidates, '\<[A-Z]+-[0-9]{2}\>', 'match')

NumberMatch = regexp(candidates, '\<(?<=[A-Z]+-)[0-9]{2}\>', 'match')

[Stephen23]

+1 neat answer.

Tip: rather than [0-9] use \d

[J. Alex Lee]

Thanks! I guess this decision depends on a few things like expression portability, and other (if any) characters to be included as "digits"

[Bhaskar R]

Thank you Alex, it worked perfectly !

but i have another situation adding another string i.e. engine number like

str = 'ENGINE-12345-23';

I have used

>> regexp(str, '[A-Z]{6}+-[0-9]{5}+-[0-9]{2}$', 'match'); % it is working for me
>>ans =

  1×1 cell array

    {'ENGINE-12345-23'}

but in the input case VCCASNSRR it should not work

>> str = 'VCCASNSRR-12345-23'
>> regexp(str, '[A-Z]{6}+-[0-9]{5}+-[0-9]{2}$', 'match')

ans =

  1×1 cell array

    {'ASNSRR-12345-23'} % this is not required it should restrict only 6 chars

How to restrict first and second string charcter to between hyphen as 6 and 5 ?

[Stephen23]

Prepend ^ to the start of the regular expression. Read the documentation to know what it does:

https://www.mathworks.com/help/matlab/matlab_prog/regular-expressions.html

[J. Alex Lee]

Stephen, do you mean prepend ^?

Bhaskar: look for "anchors" in the docs: https://www.mathworks.com/help/matlab/matlab_prog/regular-expressions.html#f0-42983

If you included \< and \> as my example shows, it should have worked too.

Your final question is unclear.

[Bhaskar R]

Sorry for if i am unclear, need to restrict the input string with specific length as

6 characters-5 digits-2 digits

Example:

str = 'ENGINE-12345-23';

In reguler expression

'...-[0-9]{2}$'; % which is restricting the last input after hyphen ( here 23) is ok
% in the same way first two string inputs(ENGINE, 12345) need restrict to 6 and 5 respectively

In the suggested link Anchors working with specfic characters but not with number of characters.

[Stephen23]

@J. Alex Lee: yes thank you, fixed now.

@Bhaskar R: use an anchor. It worked for me:

regexp(str, '^[A-Z]{6}-\d{5}-\d{2}$', 'match', 'once')

[Bhaskar R]

Yes sir, It worked perfectly

Thank you :-)

Answer 2

[dpb]

I can get you closer, but not exactly right...

regexp(s,'-\d{2}\>','match')

will return only matches of exactly two digits after the "-", but for those cases that do match, also returns the leading "-"

I'm not sure how to prevent that...I'm pretty lame at regular expressions.

[J. Alex Lee]

maybe I misunderstood the requirements. Do you need just the digits, or also the string and hyphen preceding?

candidates = [...
	"SOMETHING-2",...
	"ENGINE-45",...
	"ACCCAT-455",...
	"VCCASNSRR-12344"]

origmatches = regexp(candidates, '[A-Z]+-[0-9]{2}', 'match')
fullmatches = regexp(candidates, '[A-Z]+-[0-9]{2}$', 'match')

numericmatches = regexp(candidates,'-\d{2}\>','match')

If you want something like the last, and don't want the hyphen, i think you could use some lookbehind

[dpb]

Hmmm...rereading I see where you're coming from...looks like maybe he is trying to extract the strings+digits but only those with two digits after the hyphen. For that I'd probably just use a count if it is exactly two and forget the overhead of regexp entirely.

>> candidates(strlength(extractAfter(candidates,'-'))==2)
ans = 
    "ENGINE-45"
>> 

[J. Alex Lee]

nice solution, but still assumes what I assumed with regex, that the numbers are terminal

[dpb]

Oh, no claims it's anything other than the same thing; just don't need to figure out regex that way... :)

I've seen comments that regex is also fairly expensive; how it would compare on large array I've no idea, didn't test.

[J. Alex Lee]

yep, didn't mean to suggest otherwise either.

I don't know relative computational cost either, and hadn't seen comments about it; will definitely keep in mind. But I suppose in any case it makes sense to reach for regex only when you're at the limit of simple text processing.

[dpb]

My thinking for past 40 years! I like the quip about only aliens...am in full agreement! :)

Some people when confronted with a problem think "I know, I'll use regular expressions." Now they have two problems.

[Bhaskar R]

Thank for the for the response dpb, I have data varying with another string so I had to stick to use only reguler expressions only as of now.