feat: add update_field_metadata to edit per-field metadata (#3482)

### Summary
Adds update_field_metadata to the client SDK (Rust core, Python, and
TypeScript) so clients can edit per-field (column) Arrow metadata
(schema.fields[].metadata)

### Testing
- added unit tests
- ran E2E against a local server on both local and remote tables (set →
merge → delete), across Python sync/async and TypeScript

### Next steps
- deprecate replace_field_metadata in the python lancedb favor of this
(typescript didn't have replace_field_metadata method). This matches
Lance's API direction (Lance already deprecated replace_field_metadata
for update_field_metadata)
This commit is contained in:
Brendan Clement
2026-06-02 07:00:00 -07:00
committed by GitHub
parent 7b874905fd
commit d065be0474
17 changed files with 537 additions and 7 deletions

View File

@@ -16,8 +16,8 @@ use arrow::{
pyarrow::{FromPyArrow, PyArrowType, ToPyArrow},
};
use lancedb::table::{
AddDataMode, ColumnAlteration, Duration, NewColumnTransform, OptimizeAction, OptimizeOptions,
Table as LanceDbTable,
AddDataMode, ColumnAlteration, Duration, FieldMetadataUpdate, NewColumnTransform,
OptimizeAction, OptimizeOptions, Table as LanceDbTable,
};
use pyo3::{
Bound, FromPyObject, Py, PyAny, PyRef, PyResult, Python,
@@ -357,6 +357,27 @@ impl From<lancedb::table::AlterColumnsResult> for AlterColumnsResult {
}
}
#[pyclass(get_all, from_py_object)]
#[derive(Clone, Debug)]
pub struct UpdateFieldMetadataResult {
pub version: u64,
}
#[pymethods]
impl UpdateFieldMetadataResult {
pub fn __repr__(&self) -> String {
format!("UpdateFieldMetadataResult(version={})", self.version)
}
}
impl From<lancedb::table::UpdateFieldMetadataResult> for UpdateFieldMetadataResult {
fn from(result: lancedb::table::UpdateFieldMetadataResult) -> Self {
Self {
version: result.version,
}
}
}
#[pyclass(get_all, from_py_object)]
#[derive(Clone, Debug)]
pub struct DropColumnsResult {
@@ -1127,6 +1148,45 @@ impl Table {
Ok(())
})
}
pub fn update_field_metadata<'a>(
self_: PyRef<'a, Self>,
updates: Vec<Bound<PyDict>>,
) -> PyResult<Bound<'a, PyAny>> {
let updates = updates
.iter()
.map(|update| {
let path: String = update
.get_item("path")?
.ok_or_else(|| PyValueError::new_err("Missing path"))?
.extract()?;
let mut field_update = FieldMetadataUpdate::new(path);
if let Some(metadata) = update.get_item("metadata")? {
let metadata_dict = metadata.cast::<PyDict>()?;
for (key, value) in metadata_dict.iter() {
let key: String = key.extract()?;
if value.is_none() {
field_update = field_update.remove(key);
} else {
field_update = field_update.set(key, value.extract::<String>()?);
}
}
}
if let Some(replace) = update.get_item("replace")?
&& replace.extract::<bool>()?
{
field_update = field_update.replace();
}
Ok(field_update)
})
.collect::<PyResult<Vec<_>>>()?;
let inner = self_.inner_ref()?.clone();
future_into_py(self_.py(), async move {
let result = inner.update_field_metadata(&updates).await.infer_error()?;
Ok(UpdateFieldMetadataResult::from(result))
})
}
}
#[derive(FromPyObject)]