How to Data Incremental Update


(yeahvip) #1

Recently I use Dgraph to store hollywood star information using spyder. I save the data and relation directly to dgraph at the beginning. But with the growth of acquisition time, I should retrieve whether the obtained data is in dgraph and get uid of this data. Is there any other good way to incrementally update data? E.g. when I have some hollywood star information in dgraph, then i know that Vin Diesel is the partner of dwayne Johnson, then i need to query the uid of diesel and johnson in dgraph separately to avoid the generation of redundant nodes, then using the uid to stock this rdf.


(Daniel Mai) #2

This is exactly how you should do this. You can query-then-mutate in a single transaction.


(yeahvip) #3

But if I want to update billions of rdfs? When i use dgraph live, RDFs processed per second can be 3000. But if I query-then-mutate, RDFs processed per second can only be hundreds of orders of magnitude. Is there any other better way to incrementally add data? In my application scenario, I started with only little data, most of the data is entered after the acquisition.