feat(catalog): Implement update_table for Hive metastore catalog #1894
Conversation
vladson
force-pushed
the
hms_update_table
branch
from
3ffd73b to
b5c3d53
Compare
vladson
force-pushed
the
hms_update_table
branch
from
b5c3d53 to
bc562b7
Compare
|
Hi team, I would appreciate it if anyone could take a look at the PR, it's my first in iceberg-rust, but I hope it is in the spirit of the project. |
liurenjie1024
left a comment
liurenjie1024
left a comment
There was a problem hiding this comment.
Thanks @vladson for this pr, just finished first round of review.
crates/catalog/hms/Cargo.toml
Outdated
| metainfo = { workspace = true } | ||
| motore-macros = { workspace = true } | ||
| volo = { workspace = true } | ||
| whoami = "1.6.1" |
There was a problem hiding this comment.
We manage dependencies in workspace's root cargo.tom.
There was a problem hiding this comment.
Hey, thank you @liurenjie1024 moved whoami to the main Cargo.
| let db_name = validate_namespace(ident.namespace())?; | ||
| let tbl_name = ident.name.clone(); | ||
|
|
||
| if self.config.props.contains_key(HMS_HIVE_LOCKS_DISABLED) { |
There was a problem hiding this comment.
I'm not a hive expert, but I would suggest to start with pessimistic version only first.
There was a problem hiding this comment.
Hey, I had some experience with Hive metastore, both in setting up and using it. The thing is that many are using optimistic locking, and the problem is that in a system with multiple clients, all must use no hive locks flow, as otherwise there is a risk of data loss.
My implementation of the environment context is in line with Trino and others.
Partial implementation of HMS traditional commit stabilising fix issues with hms lock flow refactoring for optimistic locks address format
vladson
force-pushed
the
hms_update_table
branch
from
e30fa08 to
f4072a0
Compare
|
Thank you for the review; I appreciate your effort! |
2 participants
Which issue does this PR close?
What changes are included in this PR?
update_table for HMS catalog.
The implementation is based on other iceberg implementations, Trino and Hive code bases.
Are these changes tested?
The existing tests were modified, and extensive manual tests were performed.