feat(nodejs): add order_by method to Query (#3123)

This commit is contained in:
Xin Sun
2026-05-17 13:49:08 +08:00
committed by GitHub
parent f02c4cad90
commit ab2c5adf5e
19 changed files with 696 additions and 10 deletions

View File

@@ -29,6 +29,7 @@ from lancedb.query import (
MultiMatchQuery,
PhraseQuery,
BooleanQuery,
ColumnOrdering,
Occur,
LanceFtsQueryBuilder,
)
@@ -499,6 +500,36 @@ async def test_search_fts_specify_column_async(async_table):
pass
def test_search_order_by_descending(table):
table.create_fts_index("text")
rows = (
table.search("puppy")
.order_by([ColumnOrdering(column_name="count", ascending=False)])
.limit(20)
.select(["text", "count"])
.to_list()
)
for r in rows:
assert "puppy" in r["text"]
assert sorted(rows, key=lambda x: x["count"], reverse=True) == rows
def test_search_order_by_ascending(table):
table.create_fts_index("text")
rows = (
table.search("puppy")
.order_by([ColumnOrdering(column_name="count", ascending=True)])
.limit(20)
.select(["text", "count"])
.to_list()
)
for r in rows:
assert "puppy" in r["text"]
assert sorted(rows, key=lambda x: x["count"]) == rows
def test_create_index_from_table(tmp_path, table):
table.create_fts_index("text")
df = table.search("puppy").limit(5).select(["text"]).to_pandas()

View File

@@ -25,6 +25,7 @@ from lancedb.query import (
AsyncHybridQuery,
AsyncQueryBase,
AsyncVectorQuery,
ColumnOrdering,
LanceVectorQueryBuilder,
MatchQuery,
PhraseQuery,
@@ -164,6 +165,71 @@ def test_offset(table):
assert len(results_with_offset.to_pandas()) == 1
def test_order_by_plain_query(mem_db):
table = mem_db.create_table(
"test_order_by",
pa.table(
{
"group": [1, 1, 1, 2],
"score": [None, 1.0, 1.0, 0.5],
"name": ["z", "b", "a", "c"],
}
),
)
res = (
table.search()
.order_by(
[
ColumnOrdering(column_name="group", ascending=True, nulls_first=False),
ColumnOrdering(column_name="score", ascending=True, nulls_first=True),
ColumnOrdering(column_name="name", ascending=True, nulls_first=False),
]
)
.to_arrow()
)
assert res.select(["group", "score", "name"]).to_pylist() == [
{"group": 1, "score": None, "name": "z"},
{"group": 1, "score": 1.0, "name": "a"},
{"group": 1, "score": 1.0, "name": "b"},
{"group": 2, "score": 0.5, "name": "c"},
]
@pytest.mark.asyncio
async def test_order_by_async_query(mem_db_async: AsyncConnection):
table = await mem_db_async.create_table(
"test_order_by_async",
pa.table(
{
"group": [1, 1, 1, 2],
"score": [None, 1.0, 1.0, 0.5],
"name": ["z", "b", "a", "c"],
}
),
)
res = await (
table.query()
.order_by(
[
ColumnOrdering(column_name="group", ascending=True, nulls_first=False),
ColumnOrdering(column_name="score", ascending=True, nulls_first=True),
ColumnOrdering(column_name="name", ascending=True, nulls_first=False),
]
)
.to_arrow()
)
assert res.select(["group", "score", "name"]).to_pylist() == [
{"group": 1, "score": None, "name": "z"},
{"group": 1, "score": 1.0, "name": "a"},
{"group": 1, "score": 1.0, "name": "b"},
{"group": 2, "score": 0.5, "name": "c"},
]
def test_query_builder(table):
rs = (
LanceVectorQueryBuilder(table, [0, 0], "vector")

View File

@@ -16,6 +16,7 @@ from packaging.version import Version
import lancedb
from lancedb.conftest import MockTextEmbeddingFunction
from lancedb.query import ColumnOrdering
from lancedb.remote import ClientConfig
from lancedb.remote.errors import HttpError, RetryError
import pytest
@@ -660,6 +661,18 @@ def test_query_sync_maximal():
"ef": None,
"filter": "id > 0",
"columns": ["id", "name"],
"order_by": [
{
"column_name": "score",
"ascending": False,
"nulls_first": True,
},
{
"column_name": "id",
"ascending": True,
"nulls_first": False,
},
],
"vector_column": "vector2",
"fast_search": True,
"with_row_id": True,
@@ -677,6 +690,14 @@ def test_query_sync_maximal():
.refine_factor(10)
.nprobes(5)
.where("id > 0", prefilter=True)
.order_by(
[
ColumnOrdering(
column_name="score", ascending=False, nulls_first=True
),
ColumnOrdering(column_name="id", ascending=True, nulls_first=False),
]
)
.with_row_id(True)
.select(["id", "name"])
.to_list()