I try to write a function groupByMult using Ramda that applies several groupBy functions groupBys to an array of objects input:
function groupByMult(groupBys, input) { ... }
It should return a nested object that has child properties of groupBys[0], grandchild properties of groupBys[1], grand grandchildren properties of groupBys[2] and so on. The last grand child property has as value an array of objects that belong to this group path (see expected output at the bottom).
I explain the desired behaviour with an example:
My input is an array of objects. In this example all objects have a property g1, g2 and g3.
const input = [
{ g1: 'g1a', g2: 'g2a', g3: true },
{ g1: 'g1b', g2: 'g2b', g3: 42 },
{ g1: 'g1a', g2: 'g2a', g3: 'text' },
{ g1: 'g1a', g2: 'g2a', g3: false },
{ g1: 'g1a', g2: 'g2b', g3: 0 },
{ g1: 'g1a', g2: 'g2b', g3: 1 },
{ g1: 'g1a', g2: 'g2b', g3: true },
];
In my example, my group functions are:
const groupBys = [
R.prop('g1'),
R.prop('g2'),
R.compose(R.type, R.prop('g3'))
];
I call groupByMult like this
const outpout = groupByMult(groupBys, input);
And I expect output to deeply equal expected:
const expected = {
g1a: {
g2a: {
Boolean: [
{ g1: 'g1a', g2: 'g2a', g3: true },
{ g1: 'g1a', g2: 'g2a', g3: false },
],
String: [
{ g1: 'g1a', g2: 'g2a', g3: 'text' },
],
},
g2b: {
Number: [
{ g1: 'g1a', g2: 'g2b', g3: 0 },
{ g1: 'g1a', g2: 'g2b', g3: 1 },
],
Boolean: [
{ g1: 'g1a', g2: 'g2b', g3: true },
],
},
},
g1b: {
g2b: {
Number: [
{ g1: 'g1b', g2: 'g2b', g3: 42 },
]
},
},
}
output or respectively expected has child properties g1a, g1b, etc. of groupBys[0], grandchild properties g2a, g2b, etc. of groupBys[1], grand grandchildren properties Boolean, Number or String of groupBys[2], which have an array of objects that belong to this group path. For example, all objects of the array output.g1a.g2b.Boolean look like { g1: 'g1a', g2: 'g2b', Boolean: <boolean> } where <boolean> represents any boolean value.
How can I implement groupByMult to get the described behaviour?
This should be possible via a function that recursively maps over the resulting object values after applying R.groupBy.
groupByMany = R.curry((fns, items) =>
R.isEmpty(fns) ? items
: R.map(groupByMany(R.tail(fns)),
R.groupBy(R.head(fns), items)));
This function will take a list of grouping functions along with a list of objects and continue calling itself recursively until there are no more functions to group by.
For example:
input = [
{ g1: 'g1a', g2: 'g2a', g3: 'g3a' },
{ g1: 'g1b', g2: 'g2c', g3: 'g3d' },
{ g1: 'g1c', g2: 'g2a', g3: 'g3b' },
{ g1: 'g1a', g2: 'g2b', g3: 'g3a' },
{ g1: 'g1a', g2: 'g2b', g3: 'g3b' }
];
groupByMany([R.prop('g1'), R.prop('g2'), R.prop('g3')], input);
Results in something like:
{ "g1a": { "g2a": { "g3a": [{ "g1": "g1a", "g2": "g2a", "g3": "g3a" }] },
"g2b": { "g3a": [{ "g1": "g1a", "g2": "g2b", "g3": "g3a" }],
"g3b": [{ "g1": "g1a", "g2": "g2b", "g3": "g3b" }] } },
"g1b": { "g2c": { "g3d": [{ "g1": "g1b", "g2": "g2c", "g3": "g3d" }] } },
"g1c": { "g2a": { "g3b": [{ "g1": "g1c", "g2": "g2a", "g3": "g3b" }] } } }
It looks to me that what you want to do, for your example is something like this:
const groups = R.pipe(
R.groupBy(R.prop('g1')),
R.map(R.groupBy(R.prop('g2'))),
R.map(R.map(R.groupBy(R.compose(R.type, R.prop('g3')))))
);
based on an input like this:
[
R.prop('g1'),
R.prop('g2'),
R.compose(R.type, R.prop('g3'))
]
There is no difficulty wrapping each function in the list with a groupBy. That's just a map call. But there is some work involved in making that increasing list of maps for each element. I'm not sure if there's something more elegant -- or something more efficient -- than this recursive addMap helper, but this is what I came up with:
const groupByMults = (() => {
const addMap = fns => fns.length < 1 ? [] :
[R.head(fns)].concat(addMap(R.map(R.map, R.tail(fns))));
return R.curry((groupers, obj) => {
var fns = addMap(R.map(R.groupBy, groupers))
return R.apply(R.pipe)(fns)(obj);
});
}());
You can see this in action on the Ramda REPL
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With