162

I'm playing with the idea of making a completely JavaScript-based zip/unzip utility that anyone can access from a browser. They can just drag their zip directly into the browser and it'll let them download all the files within. They can also create new zip files by dragging individual files in.

I know it'd be better to do it serverside, but this project is just for a bit of fun.

Dragging files into the browser should be easy enough if I take advantage of the various methods available. (Gmail style)

Encoding/decoding should hopefully be fine. I've seen some as3 zip libraries so I'm sure I should be fine with that.

My issue is downloading the files at the end.

window.location = 'data:jpg/image;base64,/9j/4AAQSkZJR....' 

this works fine in Firefox but not in Chrome.

I can embed the files as images just fine in chrome using <img src="data:jpg/image;ba.." />, but the files won't necessarily be images. They could be any format.

Can anyone think of another solution or some kind of workaround?

Syscall
  • 18,131
  • 10
  • 32
  • 49
Mikee
  • 2,103
  • 2
  • 16
  • 16

13 Answers13

251

If you also want to give a suggested name to the file (instead of the default 'download') you can use the following in Chrome, Firefox and some IE versions:

function downloadURI(uri, name) {
  var link = document.createElement("a");
  link.download = name;
  link.href = uri;
  document.body.appendChild(link);
  link.click();
  document.body.removeChild(link);
  delete link;
}

And the following example shows it's use:

downloadURI("data:text/html,HelloWorld!", "helloWorld.txt");
owencm
  • 7,821
  • 5
  • 33
  • 53
109

function download(dataurl, filename) {
  const link = document.createElement("a");
  link.href = dataurl;
  link.download = filename;
  link.click();
}

download("data:text/html,HelloWorld!", "helloWorld.txt");

or:

function download(url, filename) {
  fetch(url)
    .then(response => response.blob())
    .then(blob => {
      const link = document.createElement("a");
      link.href = URL.createObjectURL(blob);
      link.download = filename;
      link.click();
  })
  .catch(console.error);
}

download("https://get.geojs.io/v1/ip/geo.json","geoip.json")
download("data:text/html,HelloWorld!", "helloWorld.txt");
xor
  • 4,274
  • 2
  • 23
  • 36
Zibri
  • 8,006
  • 3
  • 47
  • 40
  • 1
    click event on non existant element :D no need for appendchild. – Zibri Aug 27 '17 at 13:04
  • 2
    yes. you can simply call a.click() after you have created element 'a', and then set a's href. – user1709076 Jan 17 '19 at 13:21
  • in the past calling a.click() directly would not work but it worket with the event. Now they both work. – Zibri Mar 17 '19 at 09:48
  • 2
    This works in most modern browsers, but I'd note that appending to the document then removing is necessary to support some older browsers. – owencm Mar 22 '19 at 15:22
  • 4
    Your second solution should be used whenever the dataUri becomes too big (depends on the browser, but Chrome does not accept Uri of multiple megabytes in my experience). See also https://stackoverflow.com/questions/38781968/problems-downloading-big-filemax-15-mb-on-google-chrome . – neural5torm Jul 31 '19 at 09:24
52

Ideas:

  • Try a <a href="data:...." target="_blank"> (Untested)

  • Use downloadify instead of data URLs (would work for IE as well)

Pekka
  • 431,103
  • 135
  • 960
  • 1,075
  • 1
    @jcubic thanks for pointing out the dead link. The new download location seems to be https://github.com/dcneiner/Downloadify – Pekka Mar 03 '11 at 21:56
  • 1
    If you can't use Flash but are running a Java server-side component, you can use this: https://github.com/suprememoocow/databounce. It uses a Servlet to 'bounce' the data from client. It would be fairly easy to perform the same trick in other server side technologies, such as Python, ASP.NET etc – Andrew Newdigate Dec 16 '11 at 16:27
  • 1
    Is there a way to do it without flash? Data URL for all browsers but IE and some kind of ActiveX foo for IE? (This way I managed to play music without flash: HTML5 audio for all browsers but IE and ActiveX for IE.) – panzi Jan 29 '12 at 00:26
  • 1
    Downloadify expects flash player in the browser. If the user has no flash player, then file won't download – Jeevanandan J Aug 30 '16 at 09:23
30

Want to share my experience and help someone stuck on the downloads not working in Firefox and updated answer to 2014. The below snippet will work in both firefox and chrome and it will accept a filename:

  // Construct the <a> element
  var link = document.createElement("a");
  link.download = thefilename;
  // Construct the uri
  var uri = 'data:text/csv;charset=utf-8;base64,' + someb64data
  link.href = uri;
  document.body.appendChild(link);
  link.click();
  // Cleanup the DOM
  document.body.removeChild(link);
Martino Dino
  • 575
  • 6
  • 14
  • 4
    Unfortunately, it doesn't work in Safari. Safari doesn't seem to recognize the `download` attribute. Thanks anyways, this is as close as I can get at the moment. – Moritz Petersen Dec 23 '14 at 20:34
  • this is causing the window to redirect to the href value for me in IE 11. prevetDefault(); does not stop the behavior in IE – b_dubb Jun 01 '16 at 18:06
  • 1
    Thanks, also if `btoa` is not defined (e.g. in node eslinted frontend project) `const btxt = new Buffer(text).toString('base64'); const uri = 'data:text/csv;charset=utf-8;base64,' + btxt + ';'` – Ivan Borshchov Feb 15 '17 at 13:28
  • 2
    deleting the link is pointless, it goes out of scope on the next line and will be garbage collected, it's not a valid target for delete. – macdja38 Dec 03 '17 at 00:15
  • 1
    no need for document.body.appendChild... see my answer https://stackoverflow.com/a/45905238/236062 – Zibri Mar 16 '19 at 10:41
21

Here is a pure JavaScript solution I tested working in Firefox and Chrome but not in Internet Explorer:

function downloadDataUrlFromJavascript(filename, dataUrl) {

    // Construct the 'a' element
    var link = document.createElement("a");
    link.download = filename;
    link.target = "_blank";

    // Construct the URI
    link.href = dataUrl;
    document.body.appendChild(link);
    link.click();

    // Cleanup the DOM
    document.body.removeChild(link);
    delete link;
}

Cross-browser solutions found up until now:

downloadify -> Requires Flash

databounce -> Tested in IE 10 and 11, and doesn't work for me. Requires a servlet and some customization. (Incorrectly detects navigator. I had to set IE in compatibility mode to test, default charset in servlet, JavaScript options object with correct servlet path for absolute paths...) For non-IE browsers, it opens the file in the same window.

download.js -> http://danml.com/download.html Another library similar but not tested. Claims to be pure JavaScript, not requiring servlet nor Flash, but doesn't work on IE <= 9.

BBaysinger
  • 6,132
  • 11
  • 58
  • 114
Jesus M C
  • 291
  • 2
  • 7
13

There are several solutions but they depend on HTML5 and haven't been implemented completely in some browsers yet. Examples below were tested in Chrome and Firefox (partly works).

  1. Canvas example with save to file support. Just set your document.location.href to the data URI.
  2. Anchor download example. It uses <a href="your-data-uri" download="filename.txt"> to specify file name.
Viacheslav Dobromyslov
  • 3,039
  • 1
  • 30
  • 42
8

Combining answers from @owencm and @Chazt3n, this function will allow download of text from IE11, Firefox, and Chrome. (Sorry, I don't have access to Safari or Opera, but please add a comment if you try and it works.)

initiate_user_download = function(file_name, mime_type, text) {
    // Anything but IE works here
    if (undefined === window.navigator.msSaveOrOpenBlob) {
        var e = document.createElement('a');
        var href = 'data:' + mime_type + ';charset=utf-8,' + encodeURIComponent(text);
        e.setAttribute('href', href);
        e.setAttribute('download', file_name);
        document.body.appendChild(e);
        e.click();
        document.body.removeChild(e);
    }
    // IE-specific code
    else {
        var charCodeArr = new Array(text.length);
        for (var i = 0; i < text.length; ++i) {
            var charCode = text.charCodeAt(i);
            charCodeArr[i] = charCode;
        }
        var blob = new Blob([new Uint8Array(charCodeArr)], {type: mime_type});
        window.navigator.msSaveOrOpenBlob(blob, file_name);
    }
}

// Example:
initiate_user_download('data.csv', 'text/csv', 'Sample,Data,Here\n1,2,3\n');
kevinarpe
  • 19,075
  • 22
  • 119
  • 144
8

This can be solved 100% entirely with HTML alone. Just set the href attribute to "data:(mimetypeheader),(url)". For instance...

<a
    href="data:video/mp4,http://www.example.com/video.mp4"
    target="_blank"
    download="video.mp4"
>Download Video</a>

Working example: JSFiddle Demo.

Because we use a Data URL, we are allowed to set the mimetype which indicates the type of data to download. Documentation:

Data URLs are composed of four parts: a prefix (data:), a MIME type indicating the type of data, an optional base64 token if non-textual, and the data itself. (Source: MDN Web Docs: Data URLs.)

Components:

  • <a ...> : The link tag.
  • href="data:video/mp4,http://www.example.com/video.mp4" : Here we are setting the link to the a data: with a header preconfigured to video/mp4. This is followed by the header mimetype. I.E., for a .txt file, it would would be text/plain. And then a comma separates it from the link we want to download.
  • target="_blank" : This indicates a new tab should be opened, it's not essential, but it helps guide the browser to the desired behavior.
  • download: This is the name of the file you're downloading.
HoldOffHunger
  • 15,349
  • 8
  • 79
  • 115
3

If you only need to actually have a download action, like if you bind it to some button that will generate the URL on the fly when clicked (in Vue or React for example), you can do something as easy as this:

const link = document.createElement('a')
link.href = url
link.click()

In my case, the file is already properly named but you can set it thanks to filename if needed.

kissu
  • 24,311
  • 11
  • 36
  • 67
2

For anyone having issues in IE:

dataURItoBlob = function(dataURI) {
    var binary = atob(dataURI.split(',')[1]);
    var array = [];
    for(var i = 0; i < binary.length; i++) {
        array.push(binary.charCodeAt(i));
    }
    return new Blob([new Uint8Array(array)], {type: 'image/png'});
}

var blob = dataURItoBlob(uri);
window.navigator.msSaveOrOpenBlob(blob, "my-image.png");

This code was originally provided by @Yetti on this answer (separate question).

Bender the Greatest
  • 17,062
  • 18
  • 82
  • 145
Chazt3n
  • 1,601
  • 3
  • 17
  • 41
0

Your problem essentially boils down to "not all browsers will support this".

You could try a workaround and serve the unzipped files from a Flash object, but then you'd lose the JS-only purity (anyway, I'm not sure whether you currently can "drag files into browser" without some sort of Flash workaround - is that a HTML5 feature maybe?)

Piskvor left the building
  • 89,423
  • 44
  • 175
  • 222
-1
export const downloadAs = async (url: string, name: string) => {
  const blob = await axios.get(url, {
    headers: {
      'Content-Type': 'application/octet-stream',
    },
    responseType: 'blob',
  });
  const a = document.createElement('a');
  const href = window.URL.createObjectURL(blob.data);
  a.href = href;
  a.download = name;
  a.click();
};
user1929437
  • 423
  • 5
  • 13
-1

You can use a clean code solution, inform your url in a constant, and set it as param of open method instead in object window.

const url = "file url here"

window.open(url)
  • 1
    Please don't post code-only answers. Future readers will be grateful to see explained *why* it answers the question instead of having to infer it from the code. And, since this is an old question, please explain how it compliments the other answers. – Gert Arnold May 17 '22 at 19:08