deno.land / std@0.224.0 / csv / stringify.ts
stringify.ts
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323// Copyright 2018-2024 the Deno authors. All rights reserved. MIT license.// This module is browser compatible.
/** Array index or record key corresponding to a value for a data object. */export type PropertyAccessor = number | string;
/** * Column information. * * @param header Explicit column header name. If omitted, * the (final) property accessor is used for this value. * * @param prop Property accessor(s) used to access the value on the object */export type ColumnDetails = { header?: string; prop: PropertyAccessor | PropertyAccessor[];};
/** * The most essential aspect of a column is accessing the property holding the * data for that column on each object in the data array. If that member is at * the top level, `Column` can simply be a property accessor, which is either a * `string` (if it's a plain object) or a `number` (if it's an array). * * ```ts * const columns = [ * "name", * ]; * ``` * * Each property accessor will be used as the header for the column: * * | name | * | :--: | * | Deno | * * - If the required data is not at the top level (it's nested in other * objects/arrays), then a simple property accessor won't work, so an array of * them will be required. * * ```ts * const columns = [ * ["repo", "name"], * ["repo", "org"], * ]; * ``` * * When using arrays of property accessors, the header names inherit the value * of the last accessor in each array: * * | name | org | * | :--: | :------: | * | deno | denoland | * * - If a different column header is desired, then a `ColumnDetails` object type * can be used for each column: * * - **`header?: string`** is the optional value to use for the column header * name * * - **`prop: PropertyAccessor | PropertyAccessor[]`** is the property accessor * (`string` or `number`) or array of property accessors used to access the * data on each object * * ```ts * const columns = [ * "name", * { * prop: ["runsOn", 0], * header: "language 1", * }, * { * prop: ["runsOn", 1], * header: "language 2", * }, * ]; * ``` * * | name | language 1 | language 2 | * | :--: | :--------: | :--------: | * | Deno | Rust | TypeScript | */export type Column = ColumnDetails | PropertyAccessor | PropertyAccessor[];
/** An object (plain or array) */export type DataItem = Record<string, unknown> | unknown[];
/** Options for {@linkcode stringify}. */export type StringifyOptions = { /** Whether to include the row of headers or not. * * @default {true} */ headers?: boolean; /** * Delimiter used to separate values. Examples: * - `","` _comma_ * - `"\t"` _tab_ * - `"|"` _pipe_ * - etc. * * @default {","} */ separator?: string; /** * a list of instructions for how to target and transform the data for each * column of output. This is also where you can provide an explicit header * name for the column. */ columns?: Column[]; /** * Whether to add a * {@link https://en.wikipedia.org/wiki/Byte_order_mark | byte-order mark} to the * beginning of the file content. Required by software such as MS Excel to * properly display Unicode text. * * @default {false} */ bom?: boolean;};
const QUOTE = '"';const LF = "\n";const CRLF = "\r\n";const BYTE_ORDER_MARK = "\ufeff";
function getEscapedString(value: unknown, sep: string): string { if (value === undefined || value === null) return ""; let str = "";
if (typeof value === "object") str = JSON.stringify(value); else str = String(value);
// Is regex.test more performant here? If so, how to dynamically create? // https://stackoverflow.com/questions/3561493/ if (str.includes(sep) || str.includes(LF) || str.includes(QUOTE)) { return `${QUOTE}${str.replaceAll(QUOTE, `${QUOTE}${QUOTE}`)}${QUOTE}`; }
return str;}
type NormalizedColumn = Omit<ColumnDetails, "header" | "prop"> & { header: string; prop: PropertyAccessor[];};
function normalizeColumn(column: Column): NormalizedColumn { let header: NormalizedColumn["header"]; let prop: NormalizedColumn["prop"];
if (typeof column === "object") { if (Array.isArray(column)) { header = String(column[column.length - 1]); prop = column; } else { prop = Array.isArray(column.prop) ? column.prop : [column.prop]; header = typeof column.header === "string" ? column.header : String(prop[prop.length - 1]); } } else { header = String(column); prop = [column]; }
return { header, prop };}
/** Error thrown in {@linkcode stringify}. */export class StringifyError extends Error { /** Construct a new instance. */ constructor(message?: string) { super(message); this.name = "StringifyError"; }}
/** * Returns an array of values from an object using the property accessors * (and optional transform function) in each column */function getValuesFromItem( item: DataItem, normalizedColumns: NormalizedColumn[],): unknown[] { const values: unknown[] = [];
if (normalizedColumns.length) { for (const column of normalizedColumns) { let value: unknown = item;
for (const prop of column.prop) { if (typeof value !== "object" || value === null) continue; if (Array.isArray(value)) { if (typeof prop === "number") value = value[prop]; else { throw new StringifyError( 'Property accessor is not of type "number"', ); } } // I think this assertion is safe. Confirm? else value = (value as Record<string, unknown>)[prop]; }
values.push(value); } } else { if (Array.isArray(item)) { values.push(...item); } else if (typeof item === "object") { throw new StringifyError( "No property accessor function was provided for object", ); } else { values.push(item); } }
return values;}
/** * Write data using CSV encoding. * * @param data The source data to stringify. It's an array of items which are * plain objects or arrays. * * `DataItem: Record<string, unknown> | unknown[]` * * ```ts * const data = [ * { * name: "Deno", * repo: { org: "denoland", name: "deno" }, * runsOn: ["Rust", "TypeScript"], * }, * ]; * ``` * * @example * ```ts * import { * Column, * stringify, * } from "https://deno.land/std@$STD_VERSION/csv/stringify.ts"; * * type Character = { * age: number; * name: { * first: string; * last: string; * }; * }; * * const data: Character[] = [ * { * age: 70, * name: { * first: "Rick", * last: "Sanchez", * }, * }, * { * age: 14, * name: { * first: "Morty", * last: "Smith", * }, * }, * ]; * * let columns: Column[] = [ * ["name", "first"], * "age", * ]; * * console.log(stringify(data, { columns })); * // first,age * // Rick,70 * // Morty,14 * ``` */export function stringify( data: DataItem[], { headers = true, separator: sep = ",", columns = [], bom = false }: StringifyOptions = {},): string { if (sep.includes(QUOTE) || sep.includes(CRLF)) { const message = [ "Separator cannot include the following strings:", ' - U+0022: Quotation mark (")', " - U+000D U+000A: Carriage Return + Line Feed (\\r\\n)", ].join("\n"); throw new StringifyError(message); }
const normalizedColumns = columns.map(normalizeColumn); let output = "";
if (bom) { output += BYTE_ORDER_MARK; }
if (headers && normalizedColumns.length > 0) { output += normalizedColumns .map((column) => getEscapedString(column.header, sep)) .join(sep); output += CRLF; }
for (const item of data) { const values = getValuesFromItem(item, normalizedColumns); output += values .map((value) => getEscapedString(value, sep)) .join(sep); output += CRLF; }
return output;}
std
Version Info
6 months ago
