スプレッド演算子でimmutableにreduceの処理を書く

nodeでreduceを使って集計などを行う場合、スプレッド演算子を使ってimmutableに書けるのでメモとして残しておく

具体例でいうと次のような変数に対してname,dateごとにグループ化してvalueの値を足した計算結果を出したい場合

const rows = [
  { name: "foo", date: "2021-01-02", column1: "hoge", value: 2 },
  { name: "foo", date: "2021-01-02", column1: "fuga", value: 2 },
]

rowsに対してreduceでごにょごにょして計算結果を返すみたいなことはよくやるかと思うが

だいたいこの手のパターンだと

accumulator[key] = accumulator[key] + 1

みたいにやりがちだった

いきなりコードになってしまうがこんな感じに書ける（あくまで雰囲気）

type Row = {
  name: "foo" | "bar" | "baz";
  date: string;
  column1: string;
  value: number;
};

type AggregatedRow = {
  name: "foo" | "bar" | "baz";
  date: string;
  sum: number;
};

type Summary = {
  foo: Record<string, AggregatedRow>;
  bar: Record<string, AggregatedRow>;
  baz: Record<string, AggregatedRow>;
};

const rows: Row[] = [
  { name: "foo", date: "2021-01-02", column1: "hoge", value: 2 },
  { name: "bar", date: "2021-01-02", column1: "hoge", value: 2 },
  { name: "baz", date: "2021-01-02", column1: "hoge", value: 3 },
  { name: "foo", date: "2021-01-03", column1: "hoge", value: 1 },
  { name: "bar", date: "2021-01-03", column1: "hoge", value: 2 },
  { name: "baz", date: "2021-01-03", column1: "hoge", value: 1 },
  { name: "foo", date: "2021-01-02", column1: "fuga", value: 2 },
  { name: "bar", date: "2021-01-02", column1: "fuga", value: 1 },
  { name: "baz", date: "2021-01-02", column1: "fuga", value: 3 },
  { name: "foo", date: "2021-01-03", column1: "fuga", value: 6 },
  { name: "bar", date: "2021-01-03", column1: "fuga", value: 2 },
  { name: "baz", date: "2021-01-03", column1: "fuga", value: 3 },
];

const aggregateFn = (acc: Summary, current: Row): Summary => {
  const name = current.name;
  const date = current.date;
  // name,dateキーの更新用データを生成 Optional Chainingで既に該当キーに値があるか判定しあれば計算して結果をマージ、なければ生成する
  const aggregated = acc?.[name]?.[date]
    ? { ...acc[name][date], sum: acc[name][date].sum + current.value }
    : { ...{ name, date }, sum: current.value };
  // nameキーの更新用データを生成
  const updateValue: Record<string, AggregatedRow> = {
    ...acc[name],
    [date]: aggregated,
  };

  return { ...acc, [name]: updateValue };
};

const summary = rows.reduce(aggregateFn, {} as Summary);

console.log(summary);
// {
//   foo: {
//     '2021-01-02': { name: 'foo', date: '2021-01-02', sum: 4 },
//     '2021-01-03': { name: 'foo', date: '2021-01-03', sum: 7 }
//   },
//   bar: {
//     '2021-01-02': { name: 'bar', date: '2021-01-02', sum: 3 },
//     '2021-01-03': { name: 'bar', date: '2021-01-03', sum: 4 }
//   },
//   baz: {
//     '2021-01-02': { name: 'baz', date: '2021-01-02', sum: 6 },
//     '2021-01-03': { name: 'baz', date: '2021-01-03', sum: 4 }
//   }
// }