I'm looking for a way in Javascript to split a string into an array based on "starting" and "ending" separators rather than one separator, as str.split currently does.
For example, if I have this string:
const str = '{lang}_{cmp_abbrev}_{cmp_type}_{pl_abbrev}_{w}x{h}_d{dv}c{cv}'
The result of:
str.mySplit('{', '}');
would be this:
[
'{lang}',
'_',
'{cmp_abbrev}',
'_',
'{cmp_type}',
'_',
'{pl_abbrev}',
'_',
'{w}',
'x',
'{h}',
'_d',
'{dv}',
'c',
'{cv}'
]
Thus, it would take into consideration 2 characters instead of one character when determining how the string split should occur.
Regex to the rescue!
const str = '{lang}_{cmp_abbrev}_{cmp_type}_{pl_abbrev}_{w}x{h}_d{dv}c{cv}'
const values = [...str.matchAll(/\w+|\{\w+\}/g)].flat()
console.log(values)Array#split can take regular expressions with capturing groups:
'foo{bar}baz{}!'.split(/(\{.*?\})/g)
//=> ['foo', '{bar}', 'baz', '{}', '!']
Just be aware that empty strings can be generated e.g.,
'{foo}bar{baz}'.split(/(\{.*?\})/g)
//=> ['', '{foo}', 'bar', '{baz}', '']
'{foo}{bar}{baz}'.split(/(\{.*?\})/g)
//=> ['', '{foo}', '', '{bar}', '', '{baz}', '']
But that is both normal and to be expected. If these are undesirable you can filter them out:
'{foo}{bar}{baz}'.split(/(\{.*?\})/g).filter(Boolean)
//=> ['{foo}', '{bar}', '{baz}']
With your initial string we have:
'{lang}_{cmp_abbrev}_{cmp_type}_{pl_abbrev}_{w}x{h}_d{dv}c{cv}'.split(/(\{.*?\})/g).filter(Boolean)
//=> ['{lang}', '_', '{cmp_abbrev}', '_', '{cmp_type}', '_', '{pl_abbrev}', '_', '{w}', 'x', '{h}', '_d', '{dv}', 'c', '{cv}']