Latest Tutorials

As I’ve mentioned before,  GraphQL has become incredibly popular in the last few years . I consider GraphQL to be  the  most exciting thing to happen to APIs - there are a  ton  of benefits to GraphQL and, in fact, I’d go so far to say as that  GraphQL is the new REST . It’s not just me, either:  GraphQL has exploded in popularity , In the  State of JavaScript survey ,  over 60% of respondents  have mentioned they’ve heard of GraphQL and are interested in learning it. I’ve used GraphQL for years, and here’s just a few reasons why GraphQL is a  game changer  in application development today: With GraphQL, no longer will we ever need to make a request from the client for a specific set of data, but where a large amount is unnecessarily returned. Before we tell our GraphQL API how we want each field in our API to be resolved, we’ll need to tell GraphQL the  type  of the resolved values of each of the fields. This allows GraphQL to provide  descriptive  error messages when a query fails based on an incorrect value type. Since GraphQL and TypeScript are  both  strongly typed languages, they work  incredibly  well together. In fact, numerous tools exist (like the  Apollo Command Line Interface ) which allow us to  automatically  generate TypeScript definitions from a GraphQL API. Many different tools currently exist around working with GraphQL but none we’ve come across as powerful as the  Apollo Platform . Among a suite of many different tools exists the  Apollo Server  package where a production-ready and intuitive GraphQL API can be built within a Node.js application. The Apollo Client library also exists to allow  React  applications to fetch data from a GraphQL API with complete state management and in a declarative manner (with Hooks!). Tons of organizations have started to adopt GraphQL to power their large scale websites and mobile apps. Here are some recent examples of the many different organizations adopting GraphQL.

In this article, you'll learn how to create a Node CLI tool to convert a JSON file into a valid CSV file.JSON has become the most popular way to pass data around on the modern web, with almost universal support between APIs and server applications. However, the dominant data format in spreadsheet applications like Excel and Google Sheets is CSV, as this format is the tersest and easiest for users to understand and create. A common function that backend apps will need to perform, therefore, is the conversion of JSON to CSV. In this tutorial, we'll create a CLI script to convert an input JSON file to an output CSV file. If you don't need to customize the conversion process, it's quicker to use a third-party package like json-2-csv. At the end of the tutorial, I'll also show you how to use this instead of a custom converter. To get the most out of this article you should be familiar with the basics of Node.js, JavaScript, and async/await. You'll need a version of Node.js that includes fs.promises and supports async/await and also have NPM installed. I recommend Node v12 LTS which includes NPM 6. Let's now get started making a Node script that we can use to convert a JSON file to CSV from the command line. First, we'll create the source code directory, json-to-csv . Change into that and run npm init -y so that we're ready to add some NPM packages. Let's now create am example JSON file that we can work with called input.json . I've created a simple data schema with three properties: name, email, and date of birth. It'd be very handy to allow our utility to take in a file name input and file name output so that we can use it from the CLI. Here's the command we should be able to use from within the json-to-csv directory: So let's now create an index.js file and install the yargs package to handle CLI input: Inside index.js , let's require the yargs package and assign the argv property to a variable. This variable will effectively hold any CLI inputs captured by yargs. Nameless CLI arguments will be in an array at the _ property of argv . Let's grab these and assign them to obviously-named variables inputFileName and outputFileName . We'll also console log the values now to check they're working how we expect: For file operations, we're going to use the promises API of the fs package of Node.js. This will make handling files a lot easier than using the standard callbacks pattern. Let's do a destructure assignment to grab the readFile and writeFile methods which are all we'll need in this project. Let's now write a function that will parse the JSON file. Since file reading is an asynchronous process, let's make it an async function and name it parseJSONFile . This method will take the file name as an argument. In the method body, add a try / catch block. In the try , we'll create a variable file and assign to this await readFile(fileName) which will load the raw file. Next, we'll parse the contents as JSON and return it. In the catch block, we should console log the error so the user knows what's gone wrong. We should also exit the script by calling process.exit(1) which indicates to the shell that the process failed. We'll now write a method to convert the JavaScript array returned from the parseJSONFile to a CSV-compatible format. First, we're going to extract the values of each object in the array, discarding the keys. To do this, we'll map a new array where each element is itself an array of the object values. Next, we'll use the array unshift method to insert a header row to the top of the data. We'll pass to this the object keys of any one of the objects (since we assume they all have the same keys for the sake of simplicity). The last step is to convert the JavaScript object to CSV-compatible string. It's as simple as using the join method and joining each object with a newline ( \n ). We're not quite finished - CSV fields should be surrounded by quotes to escape any commas from within the string. There's an easy way to do this: It's fairly trivial now to write our CSV file now that we have a CSV string - we just need to call writeFile from an async method writeCSV . Just like in the parse method we'll include a try / catch block and exit on error. To run our CSV converter we'll add an IIFE to the bottom of the file. Within this function, we'll call each of our methods in the sequence we wrote them, passing data from one to the next. At the end, let's console log a message so the user knows the conversion process worked. Let's now try and run the CLI command using our example JSON file: It works! Here's what the output looks like: There's a fatal flaw in our script: if any CSV fields contain commas they will be made into separate fields during the conversion. Note in the below example what happens to the second field of the last row which includes a comma: To fix this, we'll need to escape any commas before the data gets passed to the arrayToCSV method, then unescape them afterward. We're going to create two methods to do this: escapeCommas and unescapeCommas . In the former, we'll use map to create a new array where comma values are replaced by a variable token . This token can be anything you like, so long as it doesn't occur in the CSV data. For this reason, I recommend something random like ~~~~ or !!!! . In the unescapeCommas method, we'll replace the token with the commas and restore the original content. Here's how we'll modify our run function to incorporate these new methods: With that done, the convertor can now handle commas in the content. Here's the real test of our CLI tool...can we import a converted sheet into Google Sheets? Yes, it works perfectly! Note I even put a comma in one of the fields to ensure the escape mechanism works. While it's good to understand the underlying mechanism of CSV conversion in case we need to customize it, in most projects, we'd probably just use a package like json-2-csv . Not only will this save us having to create the conversion functionality ourselves, but it also has a lot of additional features we haven't included like the ability to use different schema structures and delimiters. Let's now update our project to use this package instead. First, go ahead and install it on the command line: Next, let's require it in our project and assign it to a variable using destructuring: We can now modify our run function to use this package instead of our custom arrayToCSV method. Note we no longer need to escape our content either as the package will do this for us as well. With this change, run the CLI command again and you should get almost the same results. The key difference is that this package will only wrap fields in double-quotes if they need escaping as this still produces a valid CSV. So now you've learned how to create a CLI script for converting JSON to CSV using Node.js. Here's the complete script for your reference or if you've skimmed the article:

You can now read every newline book online! You can have access via one of two ways: Once you have access to any book, head over to " My Library " which you can find in the nav: Once you're inside, you can start reading. You'll have access to the code downloads, as well as ask any questions with inline commenting . Remember, you can get access to every book and guide with a newline Pro subscription - check it out here .

Working with dates in a Node app can be tricky as there are so many ways to format and display them. The APIs available in native JavaScript are way too tedious to use directly, so your best option is to use a date/time library. The best known and most flexible option is Moment.js . For this article, I'll presume you understand the basics of JavaScript and Node.js. To install Moment you should have Node and NPM installed on your machine. Any current version will do, but if you're installing from scratch, I'd recommend using the Node v12 LTS which includes NPM 6. Other than for simple use cases and one-offs, the JavaScript Date API will be too low-level and will require you to write many lines of code for what seems like a simple operation. Moment is an incredibly flexible JavaScript library that wraps the Date API giving you very convenient helper methods allowing you to easily perform tasks like: Moment can be used in either the browser or on a Node.js server. Let's begin by going to the terminal and installing Moment: With that done, we can now require the Moment library in a Node.js project: The first thing we'll do to use Moment is to create a new instance by calling the moment method. So what is a Moment instance, and what exactly has been assigned to the variable m in the snippet above? Think of the Moment instance as a wrapper object around a single, specific date. The wrapper provides a host of API methods that will allow you to manipulate or display the date. For example, we can use the add method which, as you'd expect, allows you to add a time period to a date: Note that Moment provides a fluent API , similar to jQuery. It's called "fluent" because the code will often read like a sentence. For example, read this line aloud and it should be immediately obvious what it does: Another aspect of fluent API methods is that they return the same object allowing you to chain additional methods for succinct and easy to read code. We said above that Moment wraps a single, specific date. So how do we specify the date we want to work with? Just like with the native JavaScript Date object, if we don't pass anything to the moment method when we first call it, the date associated with the instance will be the current date i.e. "now". What if we want to create a Moment instance based on some fixed date in the past or future? Then we can pass the date as a parameter to moment . There are several ways to do this depending on your requirements. You may be aware of some of the standards for formatting date strings (with unfortunate names like "ISO 8601" and "RFC 2822"). For example, my birthday formatted as an ISO 8601 string would look like this: "1982-10-25T08:00:15+10:00"; . Since these standards are designed for accurately communicating dates, you'll probably find that your database and other software will provide dates in one of these formats. If your date is formatted as either ISO 8601 or RFC 2822, Moment is able to automatically parse it. If your date string is not formatted using one of these standards, you'll need to tell Moment the format you're using. To do this, you supply a second argument to the moment method - a string of format tokens . Most date string formats are specified using format token templates. It's easiest to explain these using an example. Say we created a date in the format "1982-10-25". The format token template representing this would be "YYYY-MM-DD". If we wanted the same date it in the format "10/25/82" the template would be "MM/DD/YY". Hopefully, that example makes it clear that the format tokens are used for a unique date property e.g. "YYYY" corresponds to "1982", while "YY" is just "82". Format tokens are quite flexible and even allow us to create non-numeric values in dates like "25th October, 1982" - the format token string for this one would be "Do MMMM, YYYY" (note that including punctuation and other non-token values in the template is perfectly okay). For a complete list of date format tokens, see this section of the Moment docs . So, if you want to use non-standard formatting for a date, supply the format token template as the second argument and Moment will use this to parse your date string. For example, shorthand dates in Britain are normally formatted as "DD-MM-YYYY", while in America they're normally formatted as "MM-DD-YYYY". This means a date like the 10th of October, 2000 will be ambiguous (10-10-2000) and so Moment cannot parse it accurately without knowing the format you're using. If we provide the format token template, however, Moment knows what you're looking for: A Unix timestamp is a way to track time as a running total of seconds starting from the "epoch" time: January 1st, 1970 at UTC. For example, the 10th of October, 2000 is 971139600 seconds after the epoch time, therefore that value is the Unix timestamp representing that date. You'll often see timestamps used by operating systems as they're very easy to store and operate on. If you pass a number to the moment method it will be parsed as a Unix timestamp: Now we've learned how to parse dates with Moment. How can we display a date in a format we like? The API method for displaying a date in Moment is format and it can be called on a moment instance: As you can see in the console output, Moment will, by default, format a date using the ISO 8601 format. If we want to use a different format, we can simply provide a formatted token string to the method. For example, say we wanted to display the current date in the British shorthand "DD-MM-YYYY": Here's another example showing a more reader-friendly date: While it's the most popular date library, Moment is certainly not the only option. You can always consider an alternative library if Moment doesn't quite fit your needs. One complaint developers have about Moment is that the library is quite large (20kB+). As an alternative, you might try Day.js which is a 2kB minimalist replacement for Moment.js, using an almost identical API. Another complaint is that the Moment object is mutable, which can lead to confusing code. One of the maintainers of Moment has released Luxon , a Moment-inspired library that provides an immutable and unambiguous API. So now you know how to format dates using Moment and Node. You understand how to work with the current date, as well as how to use standard and non-standard date strings. You also know how to display a date in whatever format you like by supplying a format token template. Here's a snippet summarizing the main use case: Parsing and displaying dates just scratches the surface of Moment's features. You can also use it to manipulate dates, calculate durations, convert between timezones, and more. If you'd like to learn more about these features I'd recommend you start with the Moment docs:

Today we deployed a new feature: rich comments, which pairs with our earlier release of inline commenting . With this new release, you get the full power of our blog editor packed into the comments section. This means you can add simple formatting like bold , italics , and monospace to your comments, as well as more powerful features like code blocks and embeds! One of the best features of learning on newline is that you can get help inline with your learning . One of the biggest problems we've had getting help on other sites is that the comment section is completely separate from the teaching . With newline, we've added the ability to ask questions (and help others) with discussions that are inline with the teaching. This happens in two ways: If you're watching a video and you want more information, hit the "Ask a Question" button and the video will pause and you can ask your question, which is linked that that timecode. As you (and others) are watching the video, you'll see your discussion pop up in the timeline. Other students and teachers can then reply to your question. Gone are the days where you have to scroll all the way to the comments section to ask questions. On newline, you can highlight the text and post your questions (and answers) right where it's relevant: The cool thing is, these discussions have the full power of our blog editor, which means you can post syntax-highlighted code snippets, images, and even embeds like sandboxes: All of this adds up to mean that as you're learning on newline you can get help - and help others - directly alongside your learning. Try it out and let me know what you think! -- Nate and the \newline team