Skip to content
Deepbox
47advanced

DataFrame IO & Styling

Reads and writes a small sales table as CSV, JSON, XLSX and Parquet, parses dates with toDatetime, formats the table with df.style as HTML and ANSI text, and draws a line chart with df.plot.line().

What you will learn

  • Use deepbox/dataframe for DataFrame, dateRange, toDatetime, readXlsx, writeXlsx, readParquet, writeParquet, style, plot.
  • Reads and writes a small sales table as CSV, JSON, XLSX and Parquet, parses dates with toDatetime, formats the table with df.style as HTML and ANSI text, and draws a line chart with df.plot.line().

Source

index.ts
/** * Example 47: DataFrame IO & Styling * * DataFrame input and output and report formatting: CSV, JSON, XLSX and Parquet * round trips, date parsing, the string accessor, df.style (HTML and ANSI output), * and df.plot for an SVG line chart. Files are written to the output folder. */ import { mkdir, writeFile } from "node:fs/promises";import {  DataFrame,  dateRange,  readParquet,  readXlsx,  toDatetime,  writeParquet,  writeXlsx,} from "deepbox/dataframe"; const OUTPUT_DIR = "docs/examples/47-dataframe-io-styling/output"; console.log("=".repeat(60));console.log("Example 47: DataFrame IO & Styling");console.log("=".repeat(60)); await mkdir(OUTPUT_DIR, { recursive: true }); const orderDates = dateRange("2026-03-01", 6, "D");const orderDateStrings = Array.from(orderDates.data, (value) =>  value instanceof Date ? value.toISOString().slice(0, 10) : ""); const rows = [  {    orderDate: orderDateStrings[0] ?? "2026-03-01",    region: "GCC",    segment: "Enterprise",    revenue: 180_000,    supportTickets: 12,    activeCampaign: true,  },  {    orderDate: orderDateStrings[1] ?? "2026-03-02",    region: "North Africa",    segment: "SMB",    revenue: 124_000,    supportTickets: 21,    activeCampaign: false,  },  {    orderDate: orderDateStrings[2] ?? "2026-03-03",    region: "GCC",    segment: "Mid-Market",    revenue: 155_000,    supportTickets: 16,    activeCampaign: true,  },  {    orderDate: orderDateStrings[3] ?? "2026-03-04",    region: "Levant",    segment: "Enterprise",    revenue: 212_000,    supportTickets: 9,    activeCampaign: true,  },  {    orderDate: orderDateStrings[4] ?? "2026-03-05",    region: "North Africa",    segment: "Enterprise",    revenue: 167_000,    supportTickets: 13,    activeCampaign: false,  },  {    orderDate: orderDateStrings[5] ?? "2026-03-06",    region: "Levant",    segment: "SMB",    revenue: 98_000,    supportTickets: 24,    activeCampaign: false,  },]; const df = new DataFrame({  orderDate: rows.map((row) => row.orderDate),  region: rows.map((row) => row.region),  segment: rows.map((row) => row.segment),  revenue: rows.map((row) => row.revenue),  supportTickets: rows.map((row) => row.supportTickets),  activeCampaign: rows.map((row) => row.activeCampaign),}); // ============================================================================// Part 1: CSV round trip and date parsing// ============================================================================console.log("\nPart 1: CSV and Date Parsing");console.log("-".repeat(60)); const csvRoundTrip = DataFrame.fromCsvString(df.toCsvString());const parsedDates = toDatetime(rows.map((row) => row.orderDate)); console.log(  `Round-trip CSV shape: ${csvRoundTrip.shape[0]} rows x ${csvRoundTrip.shape[1]} columns`);// dayOfWeek() numbers the days from Monday = 0 to Sunday = 6console.log(  `Parsed weekday numbers: ${Array.from(parsedDates.dt.dayOfWeek().data, (value) => String(value)).join(", ")}`); // ============================================================================// Part 2: JSON, XLSX and Parquet round trips// ============================================================================console.log("\nPart 2: JSON / XLSX / Parquet");console.log("-".repeat(60)); const jsonPath = `${OUTPUT_DIR}/regional-sales.json`;await df.toJson(jsonPath);const jsonReloaded = await DataFrame.readJson(jsonPath); const xlsxBytes = writeXlsx(df.columns, rows, { sheetName: "RegionalSales" });const xlsxPath = `${OUTPUT_DIR}/regional-sales.xlsx`;await writeFile(xlsxPath, xlsxBytes);const xlsxReloaded = readXlsx(xlsxBytes); const parquetBytes = writeParquet(df.columns, rows);const parquetPath = `${OUTPUT_DIR}/regional-sales.parquet`;await writeFile(parquetPath, parquetBytes);const parquetReloaded = readParquet(parquetBytes); console.log(`JSON rows reloaded:    ${jsonReloaded.shape[0]}`);console.log(`XLSX rows reloaded:    ${xlsxReloaded.data.length}`);console.log(`Parquet rows reloaded: ${parquetReloaded.data.length}`); // ============================================================================// Part 3: String accessor and styling// ============================================================================console.log("\nPart 3: String Accessor and Styling");console.log("-".repeat(60)); const regionSeries = df.get("region");console.log(`Upper-case regions:\n${regionSeries.str.upper().toString()}`); // df.style builds a report by chaining formatting calls. The method names// highlightMax, highlightMin and backgroundGradient follow pandas.const styledHtml = df.style  .setCaption("Daily Revenue Operations Snapshot")  .highlightMax({ backgroundColor: "#dcfce7", fontWeight: "bold" })  .backgroundGradient("#fff7ed", "#fb923c")  .format("revenue", (value) => `$${Number(value).toLocaleString("en-US")}`)  .format("supportTickets", (value) => `${value} tickets`)  .toHTML(); const ansiTable = df.style  .highlightMin({ backgroundColor: "#fee2e2" })  .format("revenue", (value) => `$${Number(value).toLocaleString("en-US")}`)  .toANSI(); const htmlPath = `${OUTPUT_DIR}/styled-sales-report.html`;const ansiPath = `${OUTPUT_DIR}/styled-sales-report.txt`;await writeFile(htmlPath, styledHtml, "utf-8");await writeFile(ansiPath, ansiTable, "utf-8"); console.log(`Styled HTML report: ${htmlPath}`);console.log(`Styled ANSI table:  ${ansiPath}`); // ============================================================================// Part 4: Plot accessor// ============================================================================console.log("\nPart 4: Plot Accessor");console.log("-".repeat(60)); const trend = new DataFrame({  day: [1, 2, 3, 4, 5, 6],  revenue: rows.map((row) => row.revenue),  supportTickets: rows.map((row) => row.supportTickets),}); const trendFigure = trend.plot.line({  x: "day",  y: ["revenue", "supportTickets"],  title: "Revenue vs Support Load",  figsize: [800, 480],}); const svgPath = `${OUTPUT_DIR}/revenue-support-trend.svg`;await writeFile(svgPath, trendFigure.renderSVG().svg, "utf-8"); console.log(`Plot saved to: ${svgPath}`); // ============================================================================// Summary// ============================================================================console.log("\nKey Takeaways");console.log("-".repeat(60));console.log(  "• JSON files are read and written by DataFrame. XLSX and Parquet have built-in codecs with no extra dependency.");console.log(  "• toDatetime and dateRange build date columns, and the .dt and .str accessors clean them");console.log("• df.style writes the same table as HTML or as an ANSI string for a terminal");console.log("• df.plot.line() returns a Figure, and renderSVG() gives the SVG text"); console.log("\nDataFrame IO & Styling Example Complete!");console.log("=".repeat(60));

Output

Console output
Console output: the CSV round-trip shape, parsed weekday numbers and the number of rows read back from each file format.
Files in `output/`: `regional-sales.json`, `regional-sales.xlsx`, `regional-sales.parquet`, `styled-sales-report.html`, `styled-sales-report.txt` and `revenue-support-trend.svg`.
The older names `date_range`, `to_datetime`, `dayofweek`, `highlight_max`, `highlight_min` and `background_gradient` still work. Use `dateRange`, `toDatetime`, `dayOfWeek`, `highlightMax`, `highlightMin` and `backgroundGradient`.