My experience with Takeout lately been an unhappy one. I have posts going back to 2013, maybe earlier... but Takeout only gives me stuff going back to early 2015. The Takeout process itself says there were "errors", but doesn't provide any useful information on the nature of those errors. ☹️
I see the same
ReplyDeleteChris Harpner - Are they known to be responsive to taggings-in like this?
ReplyDeleteYou mean Takeout gives you stuff? I tried downloading the data from one private Community that has important stuff in it and all I got was one HTML file with links to online posts.And only about 2 screens-worth of links, one to a line, for a Community that goes back 7 years.
ReplyDeleteIt's not 'data' and it's not 'downloading'!
John Douglas Porter Rarely, but more than zero.
ReplyDeleteDownloaded 23gb of data, only to find errors. Whatever Google did in their upgrade over a year ago, probably borked it (could be a timestamp problem too, because before the change we had exact time/dates, not 53m and 42 weeks ago nonsense).
ReplyDeleteJohn Douglas Porter This has been my experience in the past. I suspect Google are 1) being hit with more volume than anticipated and 2) are working to address issues.
ReplyDeleteSome years ago it took months before I could retrieve a full G+ data archive, and I ended up working with a Google engineer to help resolve the problem for some of that time. It's a well-kept secret but Google can be responsive.
Do attempt to make archives.
Do check to see that you retrieved everything you'd expected to.
Do report any errors to Google, via Feedback. It's an annoying gaping void to yell into, but apparently there are readers at the other end.
Posting publicly about your success or failure is also useful for visibility.
And for those thinking "why didn't Google anticipate the demand", I'm hearing from some sources (themselves indirect) that the G+ sunset announcement was unexpected at Google also. So there's some shifting around of priorities. Take that as rumour, though having seen similar situations in the past elsewhere, plausible.
It did nothing but error for me, but some hours later I was notified that I had ~8GB in four 2GB chunks. I haven't had time to dig through it yet.
ReplyDeleteYeah, my experience is that G+ Takeout is basically worthless.
ReplyDeleteDL Keur It is not in fact worthless.
ReplyDeleteThere are limitations. The process occasionally (and at times: frequently) fails.
But it's a useful step.
A key reason I'm advising people to try Data Takeout early is so that they gain familiarity with the process, its failure modes, and the data returned.
It's also probably going to help Google iron out some bugs and capacity planning.
The failures now are not critical. They're annoying. Ideally, they shouldn't be happening. They're confusing for a lot of people (failure usually is).
But people aren't sitting facing a four-minutes-to-midnight deadline with a disk full (or empty of) worthless bits. We're ten months from doomsday, and we're learning, and (I hope, and think) Google are listening.
So we can get this shit ironed out.
And I can assure you that having your posts, comments, photos, etc., if that's what you had intended, is a lot better than not. And that I've done just this in the past, through Google Data Takeout.
(After I learned the lesson to request JSON format.)
Some dude named Tim is writing an importer for Solid.
ReplyDeletegithub.com - solid/solid-takeout-import
Errors here too
ReplyDeleteJohn Douglas Porter I have never seen a response from Google, even on the official Google help group.
ReplyDeleteFirst archive of today was only ~4GB, and was missing a LOT of posts.
ReplyDeleteSecond archive was ~40GB, and seems to have most of it. Still need to find and read the file that lists all the errors.
Had to make some space on my harddisk to actually be able to extract it.
Worth noting btw, if you choose tgz (tarred gzip) as archive format, the max archive size is increased to 50GB per part. Ignore me. 50GB archive parts are also available for Zip. Must've mis-remembered, or mis-assumed things because the filesize dropdown resets when changing the archive format.
Downside of tgz: some archivers (for instance 7-zip on Windows) will first want to unpack the gzip part of the archive, before you can extract the tar part.
I wish to add an observation: Viewing all your posts (by viewing your profile) in the web interface also seems to have similar issues. I tried to do this several times recently, and never got all my posts in a single load. And generally it's the oldest posts which somehow get dropped on the floor. May even be the same bug.
ReplyDeleteJohn Douglas Porter AFAIK the web interface no longer has a 'recent' chronological sort order, only a 'top' algorithmic order.
ReplyDeleteOne problem I've seen in most Google products is they rely on their fuzzy logic search for almost everything in places where a strict, exact filter is required. This causes bad results.
ReplyDeletewell obviously there are places where they simply can't do this, for example anything having to do with your profile.
ReplyDeleteJohn Douglas Porter This may speak to Google's storage back-end, where older / more stale data are progressively slower/harder to access. I doubt there's tape or punch cards involved, but it may be out-of-the-way spindles, or optical, or other media.
ReplyDeleteOr just less-speedy datacenters and comms paths to them.