add_sic_description.ts

#!/usr/bin/env deno run --allow-read
/**
 * Usage:
 *   ./add_descriptions.ts sic_codes.csv codes.csv
 *
 * sic_codes.csv should contain two columns:
 *   code,description
 *
 * codes.csv should contain a list of codes, one per row.
 */

// Lookup list retrieved from https://www.gov.uk/government/publications/standard-industrial-classification-of-economic-activities-sic

import {parse} from "https://deno.land/[email protected]/csv/mod.ts";

// Ensure we have the right command-line arguments.
if (Deno.args.length < 2) {
    console.error("Usage: ./add_descriptions.ts <sic_codes_with_descriptions.csv> <codes.csv>");
    Deno.exit(1);
}

const [sicCsvPath, codesCsvPath] = Deno.args;

// Read and parse the SIC codes CSV file.
const sicCsvData = await Deno.readTextFile(sicCsvPath);
const sicRows = await parse(sicCsvData, {skipFirstRow: false}) as Array<Array<string>>;

// Build a map of code to description.
// Assuming the first column is the code and the second column is the description.
const sicMap: Record<string, string> = {};
for (const row of sicRows) {
    if (row.length >= 2) {
        const code = row[0].trim();
        const description = row[1].trim();
        sicMap[code] = description;
    }
}

// Read and parse the codes CSV file.
const codesCsvData = await Deno.readTextFile(codesCsvPath);
const codeRows = await parse(codesCsvData, {skipFirstRow: false}) as Array<Array<string>>;

// Process each code and print out the code with its description.
for (const row of codeRows) {
    // Expecting each row to contain only the code in the first column.
    const code = row[0].trim();
    const description = sicMap[code] || "Description not found";
    console.log(`${code},"${description}"`);
}