Feature request: depth: false or parseDepth: false for qs.parse

[leoyli]

Similar to the relationship arrayLimit: 0 and parseArrays: false:

qs.parse('a[0]=na', { arrayLimit: 0 }) === { a: [ 'na' ] }
qs.parse('a[0]=na', { parseArrays: false }) === { a: { '0': 'na' } }

Therefore for this:

qs.parse('a[i]=na', { depth: 1 }) === { a: { i: 'na' } }
qs.parse('a[i]=na', { depth: 0 }) === { a: { '[i]': 'na' } }

I believe in some scenario, people may want to suppress parse into any depth, i.e.:

qs.parse('a[]=na', { depth: false }) === { 'a[0]': 'na' }  // proposed behaviour

so that we can expect for any query key, its value will never be an object since they are all decoupled,

qs.parse('a=A&a[]=na', { depth: false }) === { 'a': 'A', 'a[0]': 'na' }  // proposed behaviour

Any opinion?

[ljharb]

This seems like a duplicate of #246?

[leoyli]

@ljharb I read that. It actually not a duplicated since this request added new option instead of modified the current behaviour. And Indeed depth: 0 makes no sense to me, but it is consisted with the current logic, and remains the flexibility, which could be a case for people to choose. Also I think added false option should completely turn the depth parsing off, but 0 should behave literally 0depth.

[ljharb]

qs.parse('a[i]=na', { depth: 0 }) === { a: { '[i]': 'na' } } seems weird to me - zero depth, to me, seems like it would be { 'a[0]': 'na' } .

In other words, I don't think { a: { '[i]': 'na' } } ever makes sense (although I don't want to make a breaking change if I can avoid it)

[leoyli]

+1, I thought it were working like this way { 'a[0]': 'na' } when I was trying to configuring the parser function in my express app. So I'm wondering if it should be fixed like a bug, or, like what I suggested, introducing another option and have no breaking changes.

[ljharb]

There's no tests for a depth of 0, at least. Only 1, 5000, and the default of 5.

That said, maybe it'd be better to add support for false, and then in the next semver-major, make 0 work like false.

[rentalhost]

Just to who needs that it works like this report expectative:

// Input:         { a: 1, b: { c: 2 } }
// Output (spaced): a = 1 & b[c] = 2 
// Note: using lodash.
_.mapKeys(qs.parse(qs.stringify({ data: postData }), { depth: 0 }).data, (v, key) => {
    return key.replace(/^\[(\w+)]/, '$1');
})