How to extract parts of a filename using RegExp?

[jazz-y]

I don't use RegExp often, but in most cases I can handle it.

Now I have a list of filenames from cameras in the format
_MG_4991.cr2
DSC_0986.JPG

etc., in this case, between the digital index and the file extension there can be arbitrary characters (as a rule, the photographer's comments)

I want to parse these names into parts so that I get a separate prefix specific to the camera (for example, _MG_ or DSC_ or NEF_), the numerical value of the frame number and the file extension. At the same time, I need to ignore the comment of the photographer.

I read the RegExp documentation for a long time and ended up doing this:

var init = decodeURI(etTarget.path[i].name).match(new RegExp('([\D]*)(\d+)[\D]*(\.[\D]*)'))

(where etTarget.path [i] is an element of an array of objects of the File type). But it doesn't work.

I hope there are people here who are more advanced in working with strings and I will get an answer 🙂

[Kukurykus · Accepted solution]

f=['_MG_4991CommentTwo.cr2', 'DSC_0986CommentOne.JPEG', 'NEF_1234CommentThree.nef'], obj={}
while(f.length)eval(f.shift().replace(/(\w{4}\d{4})(\w+)(\..{3,4}$)/i,'obj["$1$3"]="$2"'))

f = ['_MG_4991Comment 1.cr2', 'DSC_0986Comment 2.JPEG', 'NEF_1234Comment 3.nef'], obj = {}
while(f.length)eval(f.shift().replace(/(\w{4}\d{4})([\w ]*)(\..{3,4}$)/i,'obj["$1$3"]="$2"'))

[r-bin]

f=['_MG_4991CommentTwo.cr2', 'DSC_0986CommentOne.JPEG', 'NEF_1234CommentThree.nef'], obj={}
while(f.length)eval(f.shift().replace(/(\w{4}\d{4})(\w+)(\..{3,4}$)/i,'obj["$1$3"]="$2"'))

f = ['_MG_4991Comment 1.cr2', 'DSC_0986Comment 2.JPEG', 'NEF_1234Comment 3.nef'], obj = {}
while(f.length)eval(f.shift().replace(/(\w{4}\d{4})([\w ]*)(\..{3,4}$)/i,'obj["$1$3"]="$2"'))

By @Kukurykus

messy again

f=['_MG_4991Comment Two.cr2', 'DSC_0986CommentOne.JPEG', 'NEF_1234CommentThree.nef'], obj={}
while(f.length)eval(f.shift().replace(/(\w{4}\d{4})(\w+)(\..{3,4}$)/i,'obj["$1$3"]="$2"'))

alert(obj.toSource())

[Kukurykus]

The first version was for comments with [a-z][0-9]_, another also for spaces.

[jazz-y]

var regExp = /([\D]*)(\d+)[\D]*(\.[\D]*)/,
init = regExp.exec(decodeURI(etTarget.path[0].name));

so it seems to work too

[Kukurykus]

Yes, originally I kept your code, but then found it fails for some names so I created mine.

[r-bin]

ГХМ!!!!

try {
fls=['AAA_1234 школа 10.jpg', 'DSC_0986Comment One.JPEG', 'NEF_1234CommentThree.nef'], obj={}
while(fls.length)eval(fls.shift().replace(/(\w{4}\d{4})(\w+)(\..{3,4}$)/i,'obj["$1$3"]="$2"'))

alert(obj.toSource())
} catch (e) { alert(e); }

var s = "AAA_1234 школа 10.jpg"

var regExp = /([\D]*)(\d+)[\D]*(\.[\D]*)/
init = regExp.exec(decodeURI(s));

alert(init.toSource())

???

[jazz-y]

I have a limited selection of possible comments and filenames. In any case, now I understand the approach and I can fix everything myself.

Меня больше удивляет, что в этом случае не работает match, я ожидал что он вернет тот же результат что exec. Понятно, что дело во мне, а не в match, но не пойму где ошибка

UPD. Понял. Кавычки были лишними 🙂

[Kukurykus]

That was a standard approach covering scope of typical cases. That's not hard to improve that:

eval('AAA_1234 школа 10.jpg'.replace(/(\w{4}\d{4})([\w школа]*)(\..{3,4}$)/i,'({"$1$3":"$2"})')).toSource()

[Kukurykus]

RegEx doesn't read foreign letters, but that is no problem. Either you include those letters in the code to be read or you'll use equivalent characters to gain same goal. There will be alike issue with characters from polish alphabet: ęóąśłżźćń