How to erase a substring that starts with a given pattern?

ERRATUM: c consists of character vectors, not strings.

c = {'Glucose C6H12O6'; 'Benzol C6H6'}

Hello,

Assume we have a cell array with string variables:

c = {"Glucose C6H12O6"; "Benzol C6H6"}

The goal is to remove chemical formulas from the strings and to get

c = {"Glucose"; "Benzol"}

The names can also be longer than one word. My idea for now is to define a pattern

pat = " C" + digitsPattern + "H" + digitsPattern;

and then to remove everything starting from that pattern till the end. I try to use the function eraseBetween, which can delete a substring between either two patterns or two positions. But my start point is a pattern, and end point - a position. So my questions are:

1. Is there a way to flag the end of the string as a pattern that will be unique for the given string?
2. How to find the starting position of a pattern in a string? We assume that there is only one pattern in each string.

A workaround is to add a unique pattern to the end of the string, and then delete everything between the two patterns (including them), but it is not ideal since I cannot be 100% sure the unique pattern is really unique.

Thank you in advance.

[Stephen23]

If you had a propoer string array then this would be much easier:

c = ["Glucose C6H12O6"; "Benzol C6H6"]

c = 2×1 string array
    "Glucose C6H12O6"
    "Benzol C6H6"

d = regexprep(c,'\s*(\w\d+)+','')

d = 2×1 string array
    "Glucose"
    "Benzol"

or even just get the data that you want:

d = regexp(c,'^\w+','once','match')

d = 2×1 string array
    "Glucose"
    "Benzol"

See "Why Do Strings in Cell Arrays Return an Error?" here:

https://www.mathworks.com/help/matlab/matlab_prog/frequently-asked-questions-about-string-arrays.html#mw_ea39fa70-6d47-4204-80b4-9abb58bbce77

[Wassili Dimitriew]

Now I am a bit embarrassed, because in fact I have a cell array of character vectors, not string ones. @Stephen23, I do not understand the syntax of your solutions yet, but I tested them, and they are not doing what I want:

c = {'Trans 4 Hydroxy L proline C5H9NO3'};
d = regexprep(c,'\s*(\w\d+)+','')

d = 1×1 cell array
    {'Trans 4 Hydroxy L prolineN'}

e = regexp(c,'^\w+','once','match')

e = 1×1 cell array
    {'Trans'}

The thing is, the formulas can contain other elements as well, and in some cases, the elements will not be capitalised: 'C49H55FeN4O6'.

[Stephen23]

Simpler than patterns and STRIP and all that jazz:

c = {'Trans 4 Hydroxy L proline C5H9NO3'};
d = regexprep(c,'\s*(\w+\d+)+','')

d = 1×1 cell array
    {'Trans 4 Hydroxy L proline'}

[dpb]

Yeah, assuming one can figure out the regex pattern... :>)

Answer 1

[Matt J]

c = ["Glucose C6H12O6"; "Benzol C6H6"];

pat = " C" + digitsPattern + "H" + digitsPattern;

extractBefore(c,pat)

ans = 2×1 string array
    "Glucose"
    "Benzol"

[Wassili Dimitriew]

The answer does not work for the initial, wrongly formulated, input (cell array of strings), but it works for the data that I actually have (cell array of character vectors). Therefore, I accept it; thank you!

[dpb]

While it is less convenient to enclose strings in a cell array, it can be dealt with if that were to actually be the case (not that I'd recomend other than switching storage)...

c = {"Glucose C6H12O6"; "Benzol C6H6"};
pat=' C'+digitsPattern+'H'+digitsPattern;
extractBefore([c{:}].',pat)

ans = 2×1 string array
    "Glucose"
    "Benzol"

or

extractBefore(cellstr(c),pat)

ans = 2×1 cell array
    {'Glucose'}
    {'Benzol' }

although NOTA BENE the differening result types.

As to the issue of possible case; there's a solution for that, too...

c = {"Glucose C6h12O6"; "Benzol c6H6"};
pat=caseInsensitivePattern(' C'+digitsPattern+'H'+digitsPattern);
extractBefore([c{:}],pat)

ans = 1×2 string array
    "Glucose"    "Benzol"

I don't know if the format of the input strings is controlled sufficiently well or not, but one other slight enhancement might be to wrap the result in a call to strip() if there were a chance of extra white space between the name and the formula--

c = cellstr(["Glucose  C6h12O6"; "Benzol c6H6"]); % revert to cellstr
pat=caseInsensitivePattern(' C'+digitsPattern+'H'+digitsPattern);
extractBefore(c,pat)

ans = 2×1 cell array
    {'Glucose '}
    {'Benzol'  }

or

strip(extractBefore(c,pat))

ans = 2×1 cell array
    {'Glucose'}
    {'Benzol' }

to avoid that extra blank at the end of the first chemical...

[Wassili Dimitriew]

@dpb, thank you for the comment. Indeed, strip() is something I need for tidying the data up.