Ever look at the data returned when using the Drive API? A files.list call, even if just returning a single file, can yield upwards of 4kb of data. Drive has a rich set of metadata about files, but chances are your application only needs a small fraction of what’s available.
One of the simplest but most effective optimizations you can make when building apps with the Drive API is limiting the amount of data returned to only those fields needed for your particular use case. The fields query parameter gives you that control, and the results can be dramatic.
A simple example of this is using the files.list call to display a list of files to a user. The naive query, https://www.googleapis.com/drive/v2/files?maxResults=100, generated more than 380kb of data when I ran it against my own corpus. But to render this list nicely, an app only needs a few bits of information -- the document title, icon & thumbnail URLs, the mime type, and of course the file ID.
Using the fields query parameter, the results can be trimmed to just the necessary fields and those needed for fetching subsequent pages of data. The optimized query is https://www.googleapis.com/drive/v2/files?maxResults=100&fields=items(iconLink%2Cid%2Ckind%2CmimeType%2CthumbnailLink%2Ctitle)%2CnextPageToken.
After modifying the query the resulting data was only 30k. That’s more than a 90% reduction in data size! Besides reducing the amount of data on the wire, these hints also enable us to further optimize how queries are processed. Not only is there less data to send, but also less time spent getting it in the first place.