tackling Google Drive

The boyfriend steps on the scale every day, records his weight on the whiteboard, and every once in a while enters those into a Drive spreadsheet, which has a little chart showing his progress.  (At 177 +-2, he's quite reasonable, but he feels tracking helps him keep it that way.) So I wrote a little database app for his phone.  It saves weights to a comma delimited text file and uploads it to Drive.  A macro in the spreadsheet grabs that file and puts the entries at the end of the sheet.

The fun (/sarcasm) part has been learning to upload or update this file on Drive.  The API is less than well documented, a lot of the code is compiled so you can't read it in the IDE, and that makes it extra painful since I'm barely versed in these sorts of things.  However, there's some nice sample text at the Developers site that show the basics in Java (see "Manage Drive Files").

Interestingly, Drive commands are not so much like a file system as a database.  One updates and inserts files instead of copying them.  Drive files have an ID, which is a separate creature from the name ("title").  Inserting a file will create a new file, and it could have the same title as another file that already exists -- it won't automatically overwrite it.  It's got a different ID, so it must be a different file.  So you have to look for the file and grab its ID, and from there you can update it, or do other stuff like delete, change its title, etc.

Drive requests do not ignore the Trash folder, either.  During testing, I deleted the uploaded file from Drive, expecting the upload to create a new one, and was frustrated for quite some time because my "new" file wasn't showing up. It had been happily updating the deleted file.

The other thing to watch out for is how much you're trying to do.  I followed the Developers sample code to list the files, and it grabs *all* the files, Trash, subfolders, everything. You have to add a query (yeah, database again) to restrict the file listing to where you want it to look, and better yet, the name.  After adding the query

request.setQ("'root' in parents and title = 'LatestWeights.txt' and trashed=false");

the whole process speeded up remarkably.