/** * Licensed to the Apache Software Foundation (ASF) under one * or more contributor license agreements. See the NOTICE file * distributed with this work for additional information * regarding copyright ownership. The ASF licenses this file * to you under the Apache License, Version 2.0 (the * "License"); you may not use this file except in compliance * with the License. You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, * software distributed under the License is distributed on an * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY * KIND, either express or implied. See the License for the * specific language governing permissions and limitations * under the License. */ import { DataRecord, DataRecordValue } from '@superset-ui/core'; export type TreeNode = { name: DataRecordValue; value: number; secondaryValue: number; groupBy: string; children?: TreeNode[]; }; function getMetricValue(datum: DataRecord, metric: string) { return typeof datum[metric] === 'number' ? (datum[metric] as number) : 0; } // Groups records by the raw value of `groupByKey`, keyed with a Map instead // of a plain object. A plain-object accumulator (e.g. lodash's `groupBy`) // has to coerce every key to a string to use it as a property name, so a SQL // NULL and the literal string "null" both collapse to the same "null" key // and get merged into a single group. A Map keeps them as distinct keys, so // null filtering stays deterministic regardless of what else is in the // column. // // `Date` values need special handling: a `Map` compares object keys by // identity, not value, so two separately-parsed `Date` instances for the // same timestamp would otherwise land in different groups. Canonicalizing // by epoch time keeps identical timestamps together while still keeping // `null` and `'null'` distinct. function groupByValue( data: DataRecord[], groupByKey: string, ): Map { const groups = new Map(); const keysByCanonicalTime = new Map(); data.forEach(datum => { const rawKey = datum[groupByKey]; let key = rawKey; if (rawKey instanceof Date) { const time = rawKey.getTime(); const existingKey = keysByCanonicalTime.get(time); if (existingKey !== undefined) { key = existingKey; } else { keysByCanonicalTime.set(time, rawKey); } } const group = groups.get(key); if (group) { group.push(datum); } else { groups.set(key, [datum]); } }); return groups; } export function treeBuilder( data: DataRecord[], groupBy: string[], metric: string, secondaryMetric?: string, filterNullNames?: boolean, ): TreeNode[] { const [curGroupBy, ...restGroupby] = groupBy; const curData = groupByValue(data, curGroupBy); const nodes: TreeNode[] = []; curData.forEach((value, name) => { if (!restGroupby.length) { value.forEach(datum => { const metricValue = getMetricValue(datum, metric); const secondaryValue = secondaryMetric ? getMetricValue(datum, secondaryMetric) : metricValue; nodes.push({ name, value: metricValue, secondaryValue, groupBy: curGroupBy, }); }); } else { // Children are already null-filtered by the recursive call, so the // parent's value/secondaryValue exclude hidden nulls. This keeps the // parent arc sized to its visible children (no empty gap). const children = treeBuilder( value, restGroupby, metric, secondaryMetric, filterNullNames, ); const metricValue = children.reduce( (prev, cur) => prev + (cur.value as number), 0, ); const secondaryValue = secondaryMetric ? children.reduce( (prev, cur) => prev + (cur.secondaryValue as number), 0, ) : metricValue; nodes.push({ name, children, value: metricValue, secondaryValue, groupBy: curGroupBy, }); } }); // Filter at every level so single-level charts and root nodes are covered, // not just nested children. A parent whose children were all null-filtered // is dropped too: keeping it would leave a zero-value arc that yields a NaN // secondaryValue/value ratio for coloring and tooltips. return filterNullNames ? nodes.filter(node => node.name !== null && node.children?.length !== 0) : nodes; }