# Type Conversion

**URL:** <https://talk.observablehq.com/t/type-conversion/4597>\
**Category:** Help\
**Created:** [February 7, 2021, 10:03pm UTC](https://talk.observablehq.com/t/type-conversion/4597 "2021-02-07T22:03:05Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![quicksilversurfer](https://yyz2.discourse-cdn.com/flex030/user_avatar/talk.observablehq.com/quicksilversurfer/32/4360_2.png) [@quicksilversurfer](https://talk.observablehq.com/u/quicksilversurfer)\
**Post date:** [February 7, 2021, 10:03pm UTC](https://talk.observablehq.com/t/type-conversion/4597/1 "2021-02-07T22:03:05Z")

</div>

I’m trying to convert a CSV file with numbers, strings, and currency data into something that I can work with. I’m having a hard time understanding how I should do this. I looked around in the documentation, but couldn’t really grasp the concepts there.

It’d be great if somebody can help me out a bit.

Here is the dataset -

`Rank,Name,Net Worth,Country,Industry\r  
1,Elon Musk,$203B,United States,Technology\r  
2,Jeff Bezos,$195B,United States,Technology\r  
3,Bill Gates,$135B,United States,Technology\r

And here is the function -  
d3.csvParse(await FileAttachment(“bloomberg\_billionaires.csv”).text(), d3.autoType)

autotype helps me with converting the rank to the number, but I can’t seem to get how to convert the net worth to a number. ☹  
Any help would be appreciated! And apologies for a noob question.

Thanks!

---

<div class="post-metadata">

**Author:** ![mbostock](https://yyz2.discourse-cdn.com/flex030/user_avatar/talk.observablehq.com/mbostock/32/9_2.png) [@mbostock](https://talk.observablehq.com/u/mbostock)\
**Post date:** [February 7, 2021, 10:26pm UTC](https://talk.observablehq.com/t/type-conversion/4597/2 "2021-02-07T22:26:27Z")

</div>

Have you found the documentation for d3.autoType? It might help.

> **[d3.autoType](https://observablehq.com/@d3/d3-autotype)**
>
> d3-dsv supports automatic type inference. When you parse a CSV file, it can optionally detect common data types such as numbers, dates and booleans, and convert values to the corresponding JavaScript type. Why is this useful? If you don’t perform any...

> **[d3/d3-dsv](https://github.com/d3/d3-dsv#autoType)**
>
> A parser and formatter for delimiter-separated values, such as CSV and TSV.

Specifically, d3.autoType only coerces to a number if the input is coercible to a number according to the JavaScript specification:

[https://www.ecma-international.org/ecma-262/9.0/index.html#sec-tonumber-applied-to-the-string-type](https://www.ecma-international.org/ecma-262/9.0/index.html#sec-tonumber-applied-to-the-string-type)

In this case `+"$203B"` evaluates to `NaN`, so it’s not coercing to a number. You’d need to write your own number parser for that. For example:

```auto
function parseDollars(input) {
  let match;
  if ((match = /^\$(\d+(\.\d+)?)B$/.exec(input))) return match[1] * 1e9;
  if ((match = /^\$(\d+(\.\d+)?)M$/.exec(input))) return match[1] * 1e6;
  throw new Error(`unexpected input: ${input}`);
}

```

Then you can say:

```auto
data = {
  const data = await FileAttachment("bloomberg_billionaires.csv").csv();
  for (const d of data) {
    d["Rank"] = +d["Rank"];
    d["Net Worth"] = parseDollars(d["Net Worth"]);
  }
  return data;
}

```

---

<div class="post-metadata">

**Author:** ![Cobus](https://yyz2.discourse-cdn.com/flex030/user_avatar/talk.observablehq.com/cobus/32/9670_2.png) [@Cobus](https://talk.observablehq.com/u/Cobus)\
**Post date:** [February 7, 2021, 10:38pm UTC](https://talk.observablehq.com/t/type-conversion/4597/3 "2021-02-07T22:38:50Z")

</div>

I think you have what you need (I was too slow to send my quick mock-up) but sharing here in case this example helps.

> **[csvParse formatting example](https://observablehq.com/@cobus/csvparse-formatting-example)**

---

<div class="post-metadata">

**Author:** ![quicksilversurfer](https://yyz2.discourse-cdn.com/flex030/user_avatar/talk.observablehq.com/quicksilversurfer/32/4360_2.png) [@quicksilversurfer](https://talk.observablehq.com/u/quicksilversurfer)\
**Post date:** [February 8, 2021, 1:07am UTC](https://talk.observablehq.com/t/type-conversion/4597/4 "2021-02-08T01:07:03Z")

</div>

Thank you so much Mike!

I did look at the autotype documentation, but was getting stuck at two things -

1. Parsing complex strings like $203B - still can’t wrap my head around how these expressions work, will read up on it
2. reference to the columns names - i wasn’t specifying them correctly and was getting NaN and undefined which threw me off.

But it’s a lot clearer now. I used the reference in the autotype documentation to write this more concisely -  
`parsedData2 = d3.csvParse(await FileAttachment("bloomberg_billionaires@1.csv").text(), ({rank, name, net_worth, country, industry}) => ({rank: +rank, name: name, net_worth: parseDollars(net_worth), country:country, industry:industry}))`

Again, thanks for the prompt response. Much appreciated!  
Also, love working in Observable, and excited to get better at d3.  
🍻

---

<div class="post-metadata">

**Author:** ![quicksilversurfer](https://yyz2.discourse-cdn.com/flex030/user_avatar/talk.observablehq.com/quicksilversurfer/32/4360_2.png) [@quicksilversurfer](https://talk.observablehq.com/u/quicksilversurfer)\
**Post date:** [February 8, 2021, 1:11am UTC](https://talk.observablehq.com/t/type-conversion/4597/5 "2021-02-08T01:11:09Z")

</div>

Thanks Cobus!

Even cooler, can use …d and make function even more concise. Perfect!

`parsedData2 = d3.csvParse(await FileAttachment("bloomberg_billionaires@1.csv").text(), d => ({...d, rank: +d.rank, net_worth: parseDollars(d.net_worth)}))`