Skip to content

perf(mem_wal): rank only prefiltered, newest memtable rows in a filtered vector search #9812

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
jackye1995 merged 3 commits into from
Oct 13, 2026
+658 −81
Merged

perf(mem_wal): rank only prefiltered, newest memtable rows in a filtered vector search #9812

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
dd29553
perf(mem_wal): rank only prefiltered, newest memtable rows in a filte…
hamersaw Oct 8, 2026
09d6061
test(mem_wal): check a prefiltered memtable vector search against exp…
hamersaw Oct 8, 2026
977ba1e
fix(mem_wal): order memtable key columns as the key index expects
hamersaw Oct 13, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 11 additions & 1 deletion rust/lance/src/dataset/mem_wal/index/pk.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@ use datafusion::common::ScalarValue;
use lance_core::{Error, Result};

use super::pk_key::{encode_pk_batch, encode_pk_tuple};
use super::{BTreeMemIndex, IndexStore, PrimaryKeyIndex, RowPosition};
use super::{BTreeMemIndex, IndexStore, MemIndex, PrimaryKeyIndex, RowPosition};

/// The column a composite key index is keyed on: the order-preserving encoded
/// tuple, stored as `Binary` so a [`BTreeMemIndex`] indexes it directly.
Expand Down Expand Up @@ -182,6 +182,16 @@ impl IndexStore {
self.pk_index.is_some()
}

/// The primary-key index's key columns, in the order its lookups take
/// values.
pub fn pk_columns(&self) -> Option<&[String]> {
match self.pk_index.as_ref()? {
PkIndex::Shared { entry, .. } => self.indexes.get(entry).map(|index| index.columns()),
PkIndex::Owned(OwnedPk::Single(index)) => Some(index.columns()),
PkIndex::Owned(OwnedPk::Composite { columns, .. }) => Some(columns),
}
}

/// Sorted `(key, row id)` batches for training the generation's on-disk
/// key index: the typed key for one column, the encoded tuple for several.
/// Positions are the flushed file's row ids.
Expand Down
Loading
Loading