Has anyone started looking at a tap for google ana...
# singer-tap-development
h
Has anyone started looking at a tap for google analytics 4? I took a brief look at the API, seems there is a lot there, but I don’t know if it makes sense to load all of it.
p
There was a thread a while back https://meltano.slack.com/archives/C01PKLU5D1R/p1675707121115389 where we talked about it. I also was starting work on getting the meltano labs variant updated but havent finished yet
h
Wow this is amazing! Somehow it didn’t pop up in my search, but this made my day.
I only have a tiny tiny GA account to test this on, it seems to work on 4 out of 7 (or 5/8?) streams. Do all the streams work nicely for you @Matt Menzenski ? There is a chance this is some kind of idiosyncrasy due to the GA account.
m
no ☹️ we see similar behavior to you
h
I found two bugs, that I was able to (at least somewhat) fix. One was that the data type “seconds” was parsed as int, but is a float. Second is that the default report definition for traffic_sources includes a number of dimensions that aren’t allowed. I assume the float error is a general issue, I don’t know about the metrics, perhaps that is related to the specific GA implementation. Should we try to do a few fixes and get this on to the hub? My working-branch: https://github.com/radbrt/tap-google-analytics-v4/tree/work
a
We are implementing GA4 at the moment, so will do a bit more testing with this tap over the coming months. Thanks for the pointer.
h
We should really hurry up and get something on the hub… as a competitive advantage. But I don’t want to maintain it! I try to stay as far away from google as I can.
a
Thanks for getting into it in anger @Henning Holgersen We have migrated our website analytics to now GA4 too and often demo with google analytics. So we’ll be using this tap soon too
h
Oh. sorry. My anger wasn’t supposed to be showing, only my distaste for google 😅. But sounds like we have a few possible maintainers then. Perhaps @pat_nadolny can coordinate an entry onto the hub. From what I can tell, the options are: • PR to the original connorflynn and use that one • PR to z3z1ma which is where I forked from and use that one • Matatika hosts something • Meltanolabs hosts something
a
😂 no worries. Sorry for my colloquial use of ‘in anger’. I meant to say ‘to use it for real’ not an opinion on whether Google anger is justified. It could be!
p
FYI I created https://github.com/meltano/hub/pull/1405. @Henning Holgersen related to https://meltano.slack.com/archives/CMN8HELB0/p1687974103844459?thread_ts=1687967794.072759&cid=CMN8HELB0 would you be interested in migrating your fork to meltanolabs as a new standalone tap-ga4 repo? That might allow us to all work on this together better without creating a ton of forks
💪 1
h
@pat_nadolny I really don’t want to spend many braincells on GA4, but I could, for instance, merge the
work
branch (that seems to work for me) into main, so you can fork it to meltanolabs and cut the cord to my repo?
p
I really don’t want to spend many braincells on GA4
Me neither haha sounds like everyone dislikes GA 😆 . I don't think we'd expect you to maintain it but it would serve as a place for everyone to post bug fixes in the same place. I'm happy to go with your plan
h
Sounds good. I merged it into main, so feel free to fork it to meltanolabs.
p
m
I’m finally getting around to testing this new variant. I’ll report back 🫡
m
h
@mindaugas_nizauskas Hard to say. I am of course biased because I helped create the meltanolabs variant and it is SDK-based and we have more control over it (if you want to make improvements, PRs are welcome). But if I were you, I would probably try out both to see what works and what the output looks like. If you do, please share your experience with them.
p
I'm also planning to do some work on the meltanolabs variant this week and get it up on the hub. I was waiting on credentials to test/dev. Theres also a discussion going on about whether we should merge these changes back into the original tap-google-analytics repo vs having a new standalone repo. My understanding is that UA properties are no longer active so it does make sense for the main tap-google-analytics to be GA4. Let me know if you have thoughts or reasons why we shouldnt do that
m
For me the most important thing is the ability to find the extractor in meltano hub entering google analytics
a
@mindaugas_nizauskas I think I can see it when I search: