←back to thread

Dolt is Git for Data

(github.com)
334 points gjvc | 2 comments | 23 Jun 22 11:04 UTC | HN request time: 0.001s | source
Show context
anyfactor ◴[23 Jun 22 13:27 UTC] No.31848797[source]▶
I use dolthub's public database to practice queries.

I gave their bounty program a shot but I quit when I was prompted to download/clone a 8gb database. So the best thing would be to have a dolthub account and develop the data from there.

I think if you do a little bit of git hacking you can create a branch with just an empty or a dummy database based on the schema, clone only that branch and make a PR to append to the original database. But I couldn't figure that out.

replies(1): >>31848892 #
ascar ◴[23 Jun 22 13:34 UTC] No.31848892[source]▶
Could you elaborate what's so bad about 8GB of test data? This is a database project after all and many things need a certain amount of data to show up.
replies(1): >>31850142 #
1. anyfactor ◴[23 Jun 22 15:07 UTC] No.31850142[source]▶
I tried downloading the data and my network connection glitched at 6gb, I was thrown an error and the processed exited. I believe I was recommended to restart the entire process again as there was no way continue the process where I left out. I didn’t try again.

I think there could be a better design to clone/download the database on local machine. I understand people don't actually download databases to their local machine but I expected a better solution.

replies(1): >>31850943 #
2. zachmu ◴[23 Jun 22 15:58 UTC] No.31850943[source]▶
Definitely. We have plans to allow clone to be a shallow operation, so that the original data stays remote and only your changes are local. Will make queries slower obviously, but that's the right trade-off for some people.