The JavaScript FileReader is a very powerful, efficient and asynchronous way to read the binary content of files or Blobs. Because it’s asynchronous, if you are doing high-volume, in-memory processing there is no guarantee as to the order in which reading events are completed. This can be a challenge if you have a requirement to associate some additional unique information with each file or Blob and persist it all the way thru to the end of the process. The good news is there is an easy way to do this using async tokens.
Using an asynchronous token means you can assign a unique Object, Number or String to each FileReader operation. Once you do that, the order in which the results are returned no longer matters. When each read operation completes you can simply retrieve the token uniquely associated with the original file or Blob. There really isn’t any magic. Here is a snippet of the coding pattern. You can test out a complete example on github.
function parse(blob,token,callback){ // Always create a new instance of FileReader every time. var reader = new FileReader(); // Attach the token as a property to the FileReader Object. reader.token = token; reader.onerror = function (event) { console.error(new Error(event.target.error.code).stack); } reader.onloadend = function (evt) { if(this.token != undefined){ // The reader operation is complete. // Now we can retrieve the unique token associated // with this instance of FileReader. callback(this.result,this.token); } }; reader.readAsBinaryString(blob); }
Note, it is a very bad practice to simply associate the FileReader result object with the token being passed into the parse() function’s closure. Because the results from the onloadend events can be returned in any order, each parsed result could end up being assigned the wrong token. This is an easy mistake to make and it can seriously corrupt your data.