How to update a large amount of data in dgraph every day

What do you mean by ES?

You can also use Upsert Block.

Yes, and use it in an Upsert Block.

I didn’t get it. How does this propagation works?

I’m not able to understand this scenario.

One can lease as many uid needs via /assign?what=uids&num=10000000000000000. If you use the tips given here, it will work normally if you do this.

see Self-managed cluster - Deploy

@MichelDiz would you be able to elaborate on the implications of leasing out so many uids at once? I assume there would have to be a tradeoff for doing so, right?

Honestly, I can’t see any tradeoff/issue. The leasing will increase just a little bit the storage usage things like 10 ~ 200 MB if you lease 300 million UIDs. That’s not even an issue. Also, the Zero would have to lease it (processing task). That’s all.

Ah, also you have to deal with millions/billions of UIDs on your end. That’s something. Unless you do the conversion hex to uid.