Find and merge duplicate users

PostHog / posthog

🦔 PostHog provides open-source product analytics, session recording, feature flagging and A/B testing that you can self-host.

https://posthog.com

Other

21.41k stars 1.28k forks source link

Find and merge duplicate users #1350

Open timgl opened 4 years ago

timgl commented 4 years ago

Is your feature request related to a problem?

We accidentally messed up some tracking which means we have duplicate users with the same email

Describe the solution you'd like

I want

A button on the /people page that says 'find and merge duplicates'
I want to be able to select a property to deduplicate on (email, for example)
Show me a table of 100 users at a time that are duplicate, with tick boxes to deduplicate them. Ideally also have a dropdown that shows the last couple of events for each user
- This should just be an action on the person.py endpoint
if a user needs to be de-duplicated delete one user and put those distinct IDs into the other user. Also merge any properties

Describe alternatives you've considered

Directly do it in python shell

Additional context

Thank you for your feature request – we love each and every one!

mariusandra commented 4 years ago

I fixed the case we had on app.posthog.com and did more work regarding this issue here: https://github.com/PostHog/posthog/pull/1362

safwanrahman commented 4 years ago

I would like to work on this. Can anyone assign it to me?

Twixes commented 4 years ago

I believe @mariusandra is already on it.

yakkomajuri commented 4 years ago

Merging #1382 made this even more apparent.

u-ashish commented 3 years ago

Is there an update on this? I was going to make the same issue as I noticed it in my cohorts. I've been manually merging them for now but gets a bit time consuming :). Happy to help help, in case someone isn't actively working on this...

yakkomajuri commented 3 years ago

Hey @u-ashish! So #2644 added the merge person modal which matches the description of this issue but not the title I guess?

If you mean having this done automatically - I don't believe anyone is working on it at this moment

u-ashish commented 3 years ago

@yakkomajuri Hi Yakko, yes I saw that! I was doing it manually that way but it became a little time consuming because of how many there were, so I figured I'd wait and see if the automated bit was being worked on. Thanks for letting me know.

kappa90 commented 1 year ago

Up! I would really need this.

Can we have a button in the frontend to search and merge persons together?

gprieto commented 1 year ago

Up! Any workaround in the meantime?

DNA-PC commented 1 year ago

We also need this, a UI option for mass-merging as mentioned by the OP would be great.

Meanwhile, I believe we could do it programmatically, https://posthog.com/docs/getting-started/identify-users#5-merging-users

muzammil360 commented 1 year ago

@tiina303 i noticed that you have been working on it. You added $merge command in #14625.

Since this "good first issue" is still open, how can i help? Looking to make an opensource contribution.

tiina303 commented 1 year ago

There isn't currently any plans in the near future for adding a UI option. I suggest using $merge_dangerously as a one-off or in a script. One can query the person's data via the API and run all kinds of analysis on it to figure out which users should be merged & make sure that looks sane before executing.