对对象数组进行排序并取 N 个元素

Question

in a NodeJS service I have an array with objects that have these properties:

• batchType: string
• batchId: string (is a hash)
• transactionId: string (it's a hash)

This array stores all transactions of different types of batches.

Basically what I need is to be able to get N items from the array, but respecting certain rules:

1. At least get 1 item from each type of batch
2. At least get 1 item for each batchId
3. Sometimes the array could have only one type of batch

This is an example of the array:

let batchTransactionsArray = [
  { batchType: 'type1', batchId: '123', transactionId: 'ffasf23' },
  { batchType: 'type1', batchId: '312', transactionId: '423' },
  { batchType: 'type1', batchId: '123', transactionId: '534' },
  { batchType: 'type1', batchId: '312', transactionId: '86' },
  { batchType: 'type2', batchId: '111', transactionId: '97' },
  { batchType: 'type1', batchId: '312', transactionId: '1945' },
  { batchType: 'type1', batchId: '123', transactionId: '79' },
  { batchType: 'type1', batchId: '312', transactionId: '79' },
  { batchType: 'type3', batchId: '425', transactionId: '1555645' },
  { batchType: 'type1', batchId: '123', transactionId: 'fg5' },
  { batchType: 'type1', batchId: '123', transactionId: 'jkh5' },
  { batchType: 'type1', batchId: '312', transactionId: '53j' },
  { batchType: 'type1', batchId: '111', transactionId: '4545' },
  { batchType: 'type2', batchId: '111', transactionId: '534l' },
  { batchType: 'type1', batchId: '111', transactionId: 'jkg435' },
  { batchType: 'type1', batchId: '111', transactionId: 'gfxg23' },
  { batchType: 'type1', batchId: '111', transactionId: '7asdt' },
  { batchType: 'type1', batchId: '222', transactionId: 'jdsa7' },
  { batchType: 'type3', batchId: '663', transactionId: '12423445' },
  { batchType: 'type1', batchId: '111', transactionId: '89saf6' },
  { batchType: 'type1', batchId: '111', transactionId: '12h3g' },
  { batchType: 'type1', batchId: '111', transactionId: '4h3k2hj' },
  { batchType: 'type3', batchId: '663', transactionId: '145' }
];

And an example of the output I need is (if I want 5 transactions from the array):

[{ batchType: 'type1', batchId: '123', transactionId: '534' },
 { batchType: 'type1', batchId: '312', transactionId: '86' },
 { batchType: 'type2', batchId: '111', transactionId: '97' },
 { batchType: 'type2', batchId: '111', transactionId: '534l' },
 { batchType: 'type3', batchId: '663', transactionId: '145' }
]

The criteria for sorting transactionIds would be random, there is no specific order to meet.

I was trying some lodash functions like groupBy and sortBy but no luck yet.

Here is a jsfiddle were I was playing with this: https://jsfiddle.net/20jh3ze7/

I really appreciate suggestions.

Answer 1

You could do something like this using lodash:

 let data = [ { batchType: 'type1', batchId: '123', transactionId: 'ffasf23' }, { batchType: 'type1', batchId: '312', transactionId: '423' }, { batchType: 'type1', batchId: '123', transactionId: '534' }, { batchType: 'type1', batchId: '312', transactionId: '86' }, { batchType: 'type2', batchId: '111', transactionId: '97' }, { batchType: 'type1', batchId: '312', transactionId: '1945' }, { batchType: 'type1', batchId: '123', transactionId: '79' }, { batchType: 'type1', batchId: '312', transactionId: '79' }, { batchType: 'type3', batchId: '425', transactionId: '1555645' }, { batchType: 'type1', batchId: '123', transactionId: 'fg5' }, { batchType: 'type1', batchId: '123', transactionId: 'jkh5' }, { batchType: 'type1', batchId: '312', transactionId: '53j' }, { batchType: 'type1', batchId: '111', transactionId: '4545' }, { batchType: 'type2', batchId: '111', transactionId: '534l' }, { batchType: 'type1', batchId: '111', transactionId: 'jkg435' }, { batchType: 'type1', batchId: '111', transactionId: 'gfxg23' }, { batchType: 'type1', batchId: '111', transactionId: '7asdt' }, { batchType: 'type1', batchId: '222', transactionId: 'jdsa7' }, { batchType: 'type3', batchId: '663', transactionId: '12423445' }, { batchType: 'type1', batchId: '111', transactionId: '89saf6' }, { batchType: 'type1', batchId: '111', transactionId: '12h3g' }, { batchType: 'type1', batchId: '111', transactionId: '4h3k2hj' }, { batchType: 'type3', batchId: '663', transactionId: '145' } ]; const customTake = (d, n) => { const roundRobinUnion = (arr) => { let res = [] while (_.flatten(arr).length) _.each(arr, x => x.length ? res.push(_.remove(x, (y, i) => i == 0)) : null) return _.flatten(res) } const groups = _(d) .orderBy(['batchType', 'batchId']) .groupBy('batchType') .mapValues(x => _.values(_.groupBy(x, 'batchId'))) .map(x => roundRobinUnion(x)) .value() return _.take(roundRobinUnion(groups), n) } console.log(customTake(data, 3)) console.log(customTake(data, 5)) console.log(customTake(data, 6)) console.log(customTake(data, 8))

 <script src="https://cdnjs.cloudflare.com/ajax/libs/lodash.js/4.17.10/lodash.min.js"></script>

The idea is to group by both batchType and batchId and think about this issue as a round robin union. You go through each array index and do an union of each element.

If you care about the end sort order you can always do another orderBy at the end etc.

Here is a simple example of the roundRobinUnion idea:

 const data = [ [1, 2, 3], [1], [5, 6] ] const roundRobinUnion = (arr) => { let res = [] while (_.flatten(arr).length) _.each(arr, x => x.length ? res.push(_.remove(x, (y, i) => i == 0)) : null) return _.flatten(res) } console.log(roundRobinUnion(data)) // [1,1,5,2,6,3]

 <script src="https://cdnjs.cloudflare.com/ajax/libs/lodash.js/4.17.10/lodash.min.js"></script>

Answer 2

You can do this in steps:

• first add unique batchType elements (this meets the first criterion)
• then add unique batchId elements (this meets the second criterion)
• fill the array with remaining elements if there's still some space left (to go up to N elements)

To check for uniqueness, you can use a Set . Since you said you're using lodash, you can use _.groupBy to group your array by batchItem and batchId to get unique elements from each group.

 let batchTransactionsArray = [ { batchType: 'type1', batchId: '123', transactionId: 'ffasf23' }, { batchType: 'type1', batchId: '312', transactionId: '423' }, { batchType: 'type1', batchId: '123', transactionId: '534' }, { batchType: 'type1', batchId: '312', transactionId: '86' }, { batchType: 'type2', batchId: '111', transactionId: '97' }, { batchType: 'type1', batchId: '312', transactionId: '1945' }, { batchType: 'type1', batchId: '123', transactionId: '79' }, { batchType: 'type1', batchId: '312', transactionId: '79' }, { batchType: 'type3', batchId: '425', transactionId: '1555645' }, { batchType: 'type1', batchId: '123', transactionId: 'fg5' }, { batchType: 'type1', batchId: '123', transactionId: 'jkh5' }, { batchType: 'type1', batchId: '312', transactionId: '53j' }, { batchType: 'type1', batchId: '111', transactionId: '4545' }, { batchType: 'type2', batchId: '111', transactionId: '534l' }, { batchType: 'type1', batchId: '111', transactionId: 'jkg435' }, { batchType: 'type1', batchId: '111', transactionId: 'gfxg23' }, { batchType: 'type1', batchId: '111', transactionId: '7asdt' }, { batchType: 'type1', batchId: '222', transactionId: 'jdsa7' }, { batchType: 'type3', batchId: '663', transactionId: '12423445' }, { batchType: 'type1', batchId: '111', transactionId: '89saf6' }, { batchType: 'type1', batchId: '111', transactionId: '12h3g' }, { batchType: 'type1', batchId: '111', transactionId: '4h3k2hj' }, { batchType: 'type3', batchId: '663', transactionId: '145' } ]; function getNItems(array, N = 5) { if (N >= array.length) return array; let batchTypeGroups = _.groupBy(array, e => e.batchType), res = new Set(), batchIds = new Set(); // add unique batchTypes for (let [batchType, elements] of Object.entries(batchTypeGroups)) { if (res.size === N) break; let d = elements.find(e => !batchIds.has(e.batchId)); if (d) { res.add(d); batchIds.add(d.batchId); } else { res.add(elements[0]); } } // add remaining unique batchIds let batchIdGroups = _.groupBy(array.filter(e => !batchIds.has(e.batchId)), e => e.batchId); for (let [batchId, elements] of Object.entries(batchIdGroups)) { if (res.size === N) break; res.add(elements[0]); } // add any remaining elements until we have N elements for (let e of array) { if (res.size === N) break; res.add(e); } return Array.from(res); } console.log(getNItems(batchTransactionsArray, 5));

 <script src="https://cdnjs.cloudflare.com/ajax/libs/lodash.js/4.17.11/lodash.min.js"></script>

Answer 3

I think that for this type of problem we don't need extra dependencies.

We can do it by plain javascript :)

 let batchTransactionsArray = [ { batchType: "type1", batchId: "123", transactionId: "ffasf23" }, { batchType: "type1", batchId: "312", transactionId: "423" }, { batchType: "type1", batchId: "123", transactionId: "534" }, { batchType: "type1", batchId: "312", transactionId: "86" }, { batchType: "type2", batchId: "111", transactionId: "97" }, { batchType: "type1", batchId: "312", transactionId: "1945" }, { batchType: "type1", batchId: "123", transactionId: "79" }, { batchType: "type1", batchId: "312", transactionId: "79" }, { batchType: "type3", batchId: "425", transactionId: "1555645" }, { batchType: "type1", batchId: "123", transactionId: "fg5" }, { batchType: "type1", batchId: "123", transactionId: "jkh5" }, { batchType: "type1", batchId: "312", transactionId: "53j" }, { batchType: "type1", batchId: "111", transactionId: "4545" }, { batchType: "type2", batchId: "111", transactionId: "534l" }, { batchType: "type1", batchId: "111", transactionId: "jkg435" }, { batchType: "type1", batchId: "111", transactionId: "gfxg23" }, { batchType: "type1", batchId: "111", transactionId: "7asdt" }, { batchType: "type1", batchId: "222", transactionId: "jdsa7" }, { batchType: "type3", batchId: "663", transactionId: "12423445" }, { batchType: "type1", batchId: "111", transactionId: "89saf6" }, { batchType: "type1", batchId: "111", transactionId: "12h3g" }, { batchType: "type1", batchId: "111", transactionId: "4h3k2hj" }, { batchType: "type3", batchId: "663", transactionId: "145" } ]; const getUniqueBatches = () => { const uniqueBatches = []; batchTransactionsArray.forEach(b => { const batchByType = uniqueBatches.find(findBatchByType(b.batchType)); const batchById = uniqueBatches.find(findBatchByBatchId(b.batchId)); if (!batchByType || !batchById) { uniqueBatches.push(b); } }); return uniqueBatches; }; const findBatchByType = batchType => { return batch => { return batch.batchType === batchType; }; }; const findBatchByBatchId = batchId => { return batch => { return batch.batchId === batchId; }; }; console.log(getUniqueBatches());