This is an automated email from the ASF dual-hosted git repository.
liurenjie1024 pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/iceberg-rust.git
The following commit(s) were added to refs/heads/main by this push:
new ed12ac41 feat(catalog): Add TableCommit::apply to help update metadata
(#1464)
ed12ac41 is described below
commit ed12ac41ab9f321be8796ee7ea1716472001f4df
Author: Shawn Chang <[email protected]>
AuthorDate: Tue Jun 24 17:11:08 2025 -0700
feat(catalog): Add TableCommit::apply to help update metadata (#1464)
## Which issue does this PR close?
- Closes https://github.com/apache/iceberg-rust/issues/1386
## What changes are included in this PR?
- Added `TableCommit::apply` to apply updates and requirements to a
table and return a table with updated metadata
- minor typo fix
## Are these changes tested?
added unit test
---
crates/iceberg/src/catalog/mod.rs | 92 ++++++++++++++++++++++++++++++++++++++-
crates/iceberg/src/error.rs | 2 +-
2 files changed, 92 insertions(+), 2 deletions(-)
diff --git a/crates/iceberg/src/catalog/mod.rs
b/crates/iceberg/src/catalog/mod.rs
index 5cfbae20..ebb9a66c 100644
--- a/crates/iceberg/src/catalog/mod.rs
+++ b/crates/iceberg/src/catalog/mod.rs
@@ -22,6 +22,7 @@ use std::fmt::{Debug, Display};
use std::future::Future;
use std::mem::take;
use std::ops::Deref;
+use std::sync::Arc;
use _serde::deserialize_snapshot;
use async_trait::async_trait;
@@ -313,6 +314,27 @@ impl TableCommit {
pub fn take_updates(&mut self) -> Vec<TableUpdate> {
take(&mut self.updates)
}
+
+ /// Applies this [`TableCommit`] to the given [`Table`] as part of a
catalog update.
+ /// Typically used by [`Catalog::update_table`] to validate requirements
and apply metadata updates.
+ ///
+ /// Returns a new [`Table`] with updated metadata,
+ /// or an error if validation or application fails.
+ pub fn apply(self, table: Table) -> Result<Table> {
+ // check requirements
+ for requirement in self.requirements {
+ requirement.check(Some(table.metadata()))?;
+ }
+
+ // apply updates to metadata builder
+ let mut metadata_builder = table.metadata().clone().into_builder(None);
+
+ for update in self.updates {
+ metadata_builder = update.apply(metadata_builder)?;
+ }
+
+ Ok(table.with_metadata(Arc::new(metadata_builder.build()?.metadata)))
+ }
}
/// TableRequirement represents a requirement for a table in the catalog.
@@ -884,12 +906,15 @@ mod _serde_set_statistics {
mod tests {
use std::collections::HashMap;
use std::fmt::Debug;
+ use std::fs::File;
+ use std::io::BufReader;
use serde::Serialize;
use serde::de::DeserializeOwned;
use uuid::uuid;
use super::ViewUpdate;
+ use crate::io::FileIOBuilder;
use crate::spec::{
BlobMetadata, FormatVersion, MAIN_BRANCH, NestedField, NullOrder,
Operation,
PartitionStatisticsFile, PrimitiveType, Schema, Snapshot,
SnapshotReference,
@@ -898,7 +923,10 @@ mod tests {
UnboundPartitionSpec, ViewFormatVersion, ViewRepresentation,
ViewRepresentations,
ViewVersion,
};
- use crate::{NamespaceIdent, TableCreation, TableIdent, TableRequirement,
TableUpdate};
+ use crate::table::Table;
+ use crate::{
+ NamespaceIdent, TableCommit, TableCreation, TableIdent,
TableRequirement, TableUpdate,
+ };
#[test]
fn test_parent_namespace() {
@@ -2111,4 +2139,66 @@ mod tests {
},
);
}
+
+ #[test]
+ fn test_table_commit() {
+ let table = {
+ let file = File::open(format!(
+ "{}/testdata/table_metadata/{}",
+ env!("CARGO_MANIFEST_DIR"),
+ "TableMetadataV2Valid.json"
+ ))
+ .unwrap();
+ let reader = BufReader::new(file);
+ let resp = serde_json::from_reader::<_,
TableMetadata>(reader).unwrap();
+
+ Table::builder()
+ .metadata(resp)
+
.metadata_location("s3://bucket/test/location/metadata/v2.json".to_string())
+ .identifier(TableIdent::from_strs(["ns1", "test1"]).unwrap())
+ .file_io(FileIOBuilder::new("memory").build().unwrap())
+ .build()
+ .unwrap()
+ };
+
+ let updates = vec![
+ TableUpdate::SetLocation {
+ location:
"s3://bucket/test/new_location/metadata/v2.json".to_string(),
+ },
+ TableUpdate::SetProperties {
+ updates: vec![
+ ("prop1".to_string(), "v1".to_string()),
+ ("prop2".to_string(), "v2".to_string()),
+ ]
+ .into_iter()
+ .collect(),
+ },
+ ];
+
+ let requirements = vec![TableRequirement::UuidMatch {
+ uuid: table.metadata().table_uuid,
+ }];
+
+ let table_commit = TableCommit::builder()
+ .ident(table.identifier().to_owned())
+ .updates(updates)
+ .requirements(requirements)
+ .build();
+
+ let updated_table = table_commit.apply(table).unwrap();
+
+ assert_eq!(
+ updated_table.metadata().properties.get("prop1").unwrap(),
+ "v1"
+ );
+ assert_eq!(
+ updated_table.metadata().properties.get("prop2").unwrap(),
+ "v2"
+ );
+
+ assert_eq!(
+ updated_table.metadata().location,
+ "s3://bucket/test/new_location/metadata/v2.json".to_string()
+ )
+ }
}
diff --git a/crates/iceberg/src/error.rs b/crates/iceberg/src/error.rs
index 37529ee6..07fd2641 100644
--- a/crates/iceberg/src/error.rs
+++ b/crates/iceberg/src/error.rs
@@ -39,7 +39,7 @@ pub enum ErrorKind {
/// Iceberg data is invalid.
///
/// This error is returned when we try to read a table from iceberg but
- /// failed to parse it's metadata or data file correctly.
+ /// failed to parse its metadata or data file correctly.
///
/// The table could be invalid or corrupted.
DataInvalid,
