Zig
Zig
HTML Table to CSV
See more HTML-to-XML/Text Examples
Demonstrates a method for converting an HTML table to a CSV file.Note: This example requires Chilkat v9.5.0.77 or greater.
Chilkat Zig Downloads
const std = @import("std");
const chilkat = @import("chilkat");
pub fn main(init: std.process.Init) !void {
const alloc = init.arena.allocator();
// This example requires the Chilkat API to have been previously unlocked.
// See Global Unlock Sample for sample code.
// First download the HTML containing the table
const http = try chilkat.Http.init();
defer http.deinit();
const bd_html = try chilkat.BinData.init();
defer bd_html.deinit();
http.quickGetBd("https://example-code.com/data/etf_table.html", bd_html) catch {
std.debug.print("{s}\n", .{try http.getLastErrorText(alloc)});
return;
};
// Convert to XML.
const htx = try chilkat.HtmlToXml.init();
defer htx.deinit();
htx.setHtmlBd(bd_html) catch {};
const sb_xml = try chilkat.StringBuilder.init();
defer sb_xml.deinit();
htx.toXmlSb(sb_xml) catch {};
const xml = try chilkat.Xml.init();
defer xml.deinit();
xml.loadSb(sb_xml, true) catch {};
// Remove attributes and sub-trees we don't need.
// (In other words, we're getting rid of clutter...)
var num_removed: i32 = xml.pruneTag("thead");
num_removed = xml.pruneAttribute("style");
num_removed = xml.pruneAttribute("class");
// Scrub the element and attribute content.
xml.scrub("ContentTrimEnds,ContentTrimInside,AttrTrimEnds,AttrTrimInside");
// Let's see what we have...
std.debug.print("{s}\n", .{try xml.getXml(alloc)});
// We have the following XML.
// Copy this XML into the online tool at Generate Parsing Code from XML
// as a starting point for accessing the data..
// <?xml version="1.0" encoding="utf-8"?>
// <root>
// <html>
// <head>
// <meta http-equiv="content-type" content="text/html; charset=UTF-8"/>
// </head>
// <body text="#000000" bgcolor="#FFFFFF">
// <div>
// <div>
// <table role="grid" data-scrollx="true" data-sortdirection="desc" data-sorton="-1"/>
// </div>
// </div>
// <div>
// <table id="topHoldingsTable" role="grid" data-scrollx="true" data-sortdirection="desc" data-sorton="-1">
// <tbody>
// <tr role="row">
// <td>
// <text>ITUB4</text>
// </td>
// <td>
// <text>ITAU UNIBANCO HOLDING PREF SA</text>
// </td>
// <td>
// <text>Financials</text>
// </td>
// <td>
// <text>Brazil</text>
// </td>
// <td>
// <text>10.94</text>
// </td>
// <td>
// <text>998,954,813.73</text>
// </td>
// </tr>
// <tr role="row">
// <td>
// <text>BBDC4</text>
// </td>
// <td>
// <text>BANCO BRADESCO PREF SA</text>
// </td>
// <td>
// <text>Financials</text>
// </td>
// <td>
// <text>Brazil</text>
// </td>
// <td>
// <text>9.01</text>
// </td>
// <td>
// <text>822,164,622.75</text>
// </td>
// </tr>
// ...
// ...
// ...
// </tbody>
// </table>
// </div>
// </body>
// </html>
// </root>
// This is the code generated by the online tool:
//
var i: i32 = 0;
var count_i: i32 = 0;
var table_role: [:0]const u8 = "";
var table_data_scrollx: [:0]const u8 = "";
var table_data_sortdirection: [:0]const u8 = "";
var table_data_sorton: [:0]const u8 = "";
var table_id: [:0]const u8 = "";
var j: i32 = 0;
var count_j: i32 = 0;
var tr_role: [:0]const u8 = "";
var k: i32 = 0;
var count_k: i32 = 0;
var text: [:0]const u8 = "";
i = 0;
count_i = xml.numChildrenHavingTag("html|body|div");
while (i < count_i) {
xml.setI(i);
table_role = try xml.chilkatPath(alloc, "html|body|div[i]|div|table|(role)");
table_data_scrollx = try xml.chilkatPath(alloc, "html|body|div[i]|div|table|(data-scrollx)");
table_data_sortdirection = try xml.chilkatPath(alloc, "html|body|div[i]|div|table|(data-sortdirection)");
table_data_sorton = try xml.chilkatPath(alloc, "html|body|div[i]|div|table|(data-sorton)");
table_id = try xml.chilkatPath(alloc, "html|body|div[i]|table|(id)");
table_role = try xml.chilkatPath(alloc, "html|body|div[i]|table|(role)");
table_data_scrollx = try xml.chilkatPath(alloc, "html|body|div[i]|table|(data-scrollx)");
table_data_sortdirection = try xml.chilkatPath(alloc, "html|body|div[i]|table|(data-sortdirection)");
table_data_sorton = try xml.chilkatPath(alloc, "html|body|div[i]|table|(data-sorton)");
j = 0;
count_j = xml.numChildrenHavingTag("html|body|div[i]|table|tbody|tr");
while (j < count_j) {
xml.setJ(j);
tr_role = try xml.chilkatPath(alloc, "html|body|div[i]|table|tbody|tr[j]|(role)");
k = 0;
count_k = xml.numChildrenHavingTag("html|body|div[i]|table|tbody|tr[j]|td");
while (k < count_k) {
xml.setK(k);
text = try xml.getChildContent(alloc, "html|body|div[i]|table|tbody|tr[j]|td[k]|text");
k = k + 1;
}
j = j + 1;
}
i = i + 1;
}
// Let's modify the above code to build the CSV.
const csv = try chilkat.Csv.init();
defer csv.deinit();
csv.setColumnName(0, "Ticker") catch {};
csv.setColumnName(1, "Name") catch {};
csv.setColumnName(2, "Sector") catch {};
csv.setColumnName(3, "Country") catch {};
csv.setColumnName(4, "Weight") catch {};
csv.setColumnName(5, "Notional Vaue") catch {};
i = 0;
count_i = xml.numChildrenHavingTag("html|body|div");
while (i < count_i) {
xml.setI(i);
j = 0;
count_j = xml.numChildrenHavingTag("html|body|div[i]|table|tbody|tr");
while (j < count_j) {
xml.setJ(j);
k = 0;
count_k = xml.numChildrenHavingTag("html|body|div[i]|table|tbody|tr[j]|td");
while (k < count_k) {
xml.setK(k);
csv.setCell(j, k, try xml.getChildContent(alloc, "html|body|div[i]|table|tbody|tr[j]|td[k]|text")) catch {};
k = k + 1;
}
j = j + 1;
}
i = i + 1;
}
csv.saveFile("qa_output/brasil_etf.csv") catch {};
const csv_str = try csv.saveToString(alloc);
std.debug.print("{s}\n", .{csv_str});
// Our CSV looks like this:
// Ticker,Name,Sector,Country,Weight,Notional Vaue
// ITUB4,ITAU UNIBANCO HOLDING PREF SA,Financials,Brazil,10.94,"998,954,813.73"
// BBDC4,BANCO BRADESCO PREF SA,Financials,Brazil,9.01,"822,164,622.75"
// VALE3,CIA VALE DO RIO DOCE SH,Materials,Brazil,8.60,"785,290,260.07"
// PETR4,PETROLEO BRASILEIRO PREF SA,Energy,Brazil,5.68,"518,124,434.10"
// PETR3,PETROBRAS,Energy,Brazil,4.86,"443,254,438.53"
// B3SA3,B3 BRASIL BOLSA BALCAO SA,Financials,Brazil,4.57,"417,636,740.16"
// ABEV3,AMBEV SA,Consumer Staples,Brazil,4.57,"417,216,913.63"
// BBAS3,BANCO DO BRASIL SA,Financials,Brazil,3.25,"296,921,232.15"
// ITSA4,ITAUSA INVESTIMENTOS ITAU PREF SA,Financials,Brazil,2.90,"265,153,684.52"
// LREN3,LOJAS RENNER SA,Consumer Discretionary,Brazil,2.25,"205,832,175.98"
//
}