Bump version: 0.25.1-beta.4 → 0.25.1

### Bug Fix: Undefined Values in Nullable Fields

**Issue**: When inserting data with `undefined` values into nullable
fields, LanceDB was incorrectly coercing them to default values (`false`
for booleans, `NaN` for numbers, `""` for strings) instead of `null`.

**Fix**: Modified the `makeVector()` function in `arrow.ts` to properly
convert `undefined` values to `null` for nullable fields before passing
data to Apache Arrow.

fixes: #2645

**Result**: Now `{ text: undefined, number: undefined, bool: undefined
}` correctly becomes `{ text: null, number: null, bool: null }` when
fields are marked as nullable in the schema.

**Files Changed**: 
- `nodejs/lancedb/arrow.ts` (core fix)
- `nodejs/__test__/arrow.test.ts` (test coverage)

- This ensures proper null handling for nullable fields as expected by
users.

---------

Co-authored-by: Will Jones <willjones127@gmail.com>

.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.22.1-beta.1"
+current_version = "0.22.1-beta.3"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

.github/workflows/pypi-publish.yml

@@ -56,7 +56,7 @@ jobs:
           pypi_token: ${{ secrets.LANCEDB_PYPI_API_TOKEN }}
           fury_token: ${{ secrets.FURY_TOKEN }}
   mac:
-    timeout-minutes: 60
+    timeout-minutes: 90
     runs-on: ${{ matrix.config.runner }}
     strategy:
       matrix:
@@ -64,7 +64,7 @@ jobs:
           - target: x86_64-apple-darwin
             runner: macos-13
           - target: aarch64-apple-darwin
-            runner: macos-14
+            runner: warp-macos-14-arm64-6x
     env:
       MACOSX_DEPLOYMENT_TARGET: 10.15
     steps:

Cargo.lock

@@ -1153,7 +1153,7 @@ dependencies = [
  "bitflags 2.9.4",
  "cexpr",
  "clang-sys",
- "itertools 0.11.0",
+ "itertools 0.12.1",
  "lazy_static",
  "lazycell",
  "log",
@@ -2929,6 +2929,18 @@ version = "0.2.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "f8eb564c5c7423d25c886fb561d1e4ee69f72354d16918afa32c08811f6b6a55"
 
+[[package]]
+name = "fastbloom"
+version = "0.14.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "18c1ddb9231d8554c2d6bdf4cfaabf0c59251658c68b6c95cd52dd0c513a912a"
+dependencies = [
+ "getrandom 0.3.3",
+ "libm",
+ "rand 0.9.2",
+ "siphasher",
+]
+
 [[package]]
 name = "fastdivide"
 version = "0.4.2"
@@ -3028,8 +3040,9 @@ checksum = "42703706b716c37f96a77aea830392ad231f44c9e9a67872fa5548707e11b11c"
 
 [[package]]
 name = "fsst"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "fe0a0b1d16ce6b863be8ab766004d89ebf0779fd6ce31b0ef3bbc7fedaaad373"
 dependencies = [
  "arrow-array",
  "rand 0.9.2",
@@ -4206,8 +4219,9 @@ dependencies = [
 
 [[package]]
 name = "lance"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "42171f2af5d377e6bbcc8a8572144ee15b73a8f78ceb6160f1adeabf0d0f3e3c"
 dependencies = [
  "arrow",
  "arrow-arith",
@@ -4270,8 +4284,9 @@ dependencies = [
 
 [[package]]
 name = "lance-arrow"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "25ef9499a1e581112f45fbf743fdc8e24830cda0bd13396b11c71aa6e6cba083"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -4289,8 +4304,9 @@ dependencies = [
 
 [[package]]
 name = "lance-bitpacking"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1101fffd5b161bbdc6e932d6c0a7f94cb1752b0f8cd6d18ef9064052ab901a84"
 dependencies = [
  "arrayref",
  "paste",
@@ -4299,8 +4315,9 @@ dependencies = [
 
 [[package]]
 name = "lance-core"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "527ee5e6472d058d8c66c702fbe318a3f60f971e652e60dcfc6349bdbc9b0733"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -4335,8 +4352,9 @@ dependencies = [
 
 [[package]]
 name = "lance-datafusion"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "65a80f7f15f2d941ec7b8253625cbb8e12081ea27584dd1fbc657fb9fb377f7a"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -4365,8 +4383,9 @@ dependencies = [
 
 [[package]]
 name = "lance-datagen"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0495c8afa18f246ac4b337c47d7827560283783963dd2177862d91161478fd79"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -4383,8 +4402,9 @@ dependencies = [
 
 [[package]]
 name = "lance-encoding"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0e80e9ae49d68b95d58e77d9177f68983dce4f0803ef42840e1631b38dd66adc"
 dependencies = [
  "arrow-arith",
  "arrow-array",
@@ -4412,6 +4432,7 @@ dependencies = [
  "prost-types",
  "rand 0.9.2",
  "snafu",
+ "strum",
  "tokio",
  "tracing",
  "xxhash-rust",
@@ -4420,8 +4441,9 @@ dependencies = [
 
 [[package]]
 name = "lance-file"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f1707f9f5097b36c82d3a8524bb41c762c80d5dfa5e32aa7bfc6a1c0847a1cce"
 dependencies = [
  "arrow-arith",
  "arrow-array",
@@ -4455,8 +4477,9 @@ dependencies = [
 
 [[package]]
 name = "lance-index"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "28ab52586a5a7f5371a5abf4862968231f8c0232ce0780bc456f1ec16e9370f9"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -4477,6 +4500,7 @@ dependencies = [
  "datafusion-sql",
  "deepsize",
  "dirs",
+ "fastbloom",
  "fst",
  "futures",
  "half",
@@ -4491,6 +4515,7 @@ dependencies = [
  "lance-io",
  "lance-linalg",
  "lance-table",
+ "libm",
  "log",
  "num-traits",
  "object_store",
@@ -4507,13 +4532,15 @@ dependencies = [
  "tempfile",
  "tokio",
  "tracing",
+ "twox-hash",
  "uuid",
 ]
 
 [[package]]
 name = "lance-io"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d606f9f6a7f8ec2cacf28dfce7b2fc39e7db9f0ec77f907b8e47c756e3dd163b"
 dependencies = [
  "arrow",
  "arrow-arith",
@@ -4553,8 +4580,9 @@ dependencies = [
 
 [[package]]
 name = "lance-linalg"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c9f1a94a5d966ff1eae817a835e3a57b34f73300f83a43bb28e7e2806695b8ba"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -4577,8 +4605,9 @@ dependencies = [
 
 [[package]]
 name = "lance-table"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "fac5c0ca6e5c285645465b95fb99fc464a1fd22a6d4b32ae0e0760f06b4b8a7f"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -4616,8 +4645,9 @@ dependencies = [
 
 [[package]]
 name = "lance-testing"
-version = "0.35.0"
+version = "0.37.0"
-source = "git+https://github.com/lancedb/lance.git?tag=v0.35.0-beta.4#e842a8f922b90c298c356dd1c6afdc83ca5253f2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "384acc1dd13379a2ae24f3e3635d9c1f4fb4dc1534f7ffd2740c268f2eb73455"
 dependencies = [
  "arrow-array",
  "arrow-schema",
@@ -4628,7 +4658,7 @@ dependencies = [
 
 [[package]]
 name = "lancedb"
-version = "0.22.1-beta.1"
+version = "0.22.1-beta.3"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -4715,7 +4745,7 @@ dependencies = [
 
 [[package]]
 name = "lancedb-nodejs"
-version = "0.22.1-beta.1"
+version = "0.22.1-beta.3"
 dependencies = [
  "arrow-array",
  "arrow-ipc",
@@ -4735,7 +4765,7 @@ dependencies = [
 
 [[package]]
 name = "lancedb-python"
-version = "0.25.1-beta.1"
+version = "0.25.1-beta.3"
 dependencies = [
  "arrow",
  "async-trait",
@@ -7781,6 +7811,15 @@ version = "0.11.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7da8b5736845d9f2fcb837ea5d9e2628564b3b043a70948a3f0b778838c5fb4f"
 
+[[package]]
+name = "strum"
+version = "0.25.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "290d54ea6f91c969195bdbcd7442c8c2a2ba87da8bf60a7ee86a235d4bc1e125"
+dependencies = [
+ "strum_macros",
+]
+
 [[package]]
 name = "strum_macros"
 version = "0.25.3"
@@ -8441,6 +8480,9 @@ name = "twox-hash"
 version = "2.1.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "8b907da542cbced5261bd3256de1b3a1bf340a3d37f93425a07362a1d687de56"
+dependencies = [
+ "rand 0.9.2",
+]
 
 [[package]]
 name = "typenum"

Cargo.toml

@@ -15,14 +15,14 @@ categories = ["database-implementations"]
 rust-version = "1.78.0"
 
 [workspace.dependencies]
-lance = { "version" = "=0.35.0", default-features = false, "features" = ["dynamodb"], "tag" = "v0.35.0-beta.4", "git" = "https://github.com/lancedb/lance.git" }
+lance = { "version" = "=0.37.0", default-features = false, "features" = ["dynamodb"] }
-lance-io = { "version" = "=0.35.0", default-features = false, "tag" = "v0.35.0-beta.4", "git" = "https://github.com/lancedb/lance.git" }
+lance-io = { "version" = "=0.37.0", default-features = false }
-lance-index = { "version" = "=0.35.0", "tag" = "v0.35.0-beta.4", "git" = "https://github.com/lancedb/lance.git" }
+lance-index = "=0.37.0"
-lance-linalg = { "version" = "=0.35.0", "tag" = "v0.35.0-beta.4", "git" = "https://github.com/lancedb/lance.git" }
+lance-linalg = "=0.37.0"
-lance-table = { "version" = "=0.35.0", "tag" = "v0.35.0-beta.4", "git" = "https://github.com/lancedb/lance.git" }
+lance-table = "=0.37.0"
-lance-testing = { "version" = "=0.35.0", "tag" = "v0.35.0-beta.4", "git" = "https://github.com/lancedb/lance.git" }
+lance-testing = "=0.37.0"
-lance-datafusion = { "version" = "=0.35.0", "tag" = "v0.35.0-beta.4", "git" = "https://github.com/lancedb/lance.git" }
+lance-datafusion = "=0.37.0"
-lance-encoding = { "version" = "=0.35.0", "tag" = "v0.35.0-beta.4", "git" = "https://github.com/lancedb/lance.git" }
+lance-encoding = "=0.37.0"
 # Note that this one does not include pyarrow
 arrow = { version = "55.1", optional = false }
 arrow-array = "55.1"

ci/set_lance_version.py

@@ -1,4 +1,5 @@
 import argparse
+import re
 import sys
 import json
 
@@ -18,8 +19,12 @@ def run_command(command: str) -> str:
 
 def get_latest_stable_version() -> str:
     version_line = run_command("cargo info lance | grep '^version:'")
-    version = version_line.split(" ")[1].strip()
+    # Example output: "version: 0.35.0 (latest 0.37.0)"
-    return version
+    match = re.search(r'\(latest ([0-9.]+)\)', version_line)
+    if match:
+        return match.group(1)
+    # Fallback: use the first version after 'version:'
+    return version_line.split("version:")[1].split()[0].strip()
 
 
 def get_latest_preview_version() -> str:

java/core/pom.xml

@@ -8,7 +8,7 @@
     <parent>
         <groupId>com.lancedb</groupId>
         <artifactId>lancedb-parent</artifactId>
-        <version>0.22.1-beta.1</version>
+        <version>0.22.1-beta.3</version>
         <relativePath>../pom.xml</relativePath>
     </parent>
 

java/lance-namespace/pom.xml

@@ -8,7 +8,7 @@
     <parent>
         <groupId>com.lancedb</groupId>
         <artifactId>lancedb-parent</artifactId>
-        <version>0.22.1-beta.1</version>
+        <version>0.22.1-beta.3</version>
         <relativePath>../pom.xml</relativePath>
     </parent>
 

java/pom.xml

@@ -6,7 +6,7 @@
 
     <groupId>com.lancedb</groupId>
     <artifactId>lancedb-parent</artifactId>
-    <version>0.22.1-beta.1</version>
+    <version>0.22.1-beta.3</version>
     <packaging>pom</packaging>
     <name>${project.artifactId}</name>
     <description>LanceDB Java SDK Parent POM</description>

nodejs/Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "lancedb-nodejs"
 edition.workspace = true
-version = "0.22.1-beta.1"
+version = "0.22.1-beta.3"
 license.workspace = true
 description.workspace = true
 repository.workspace = true

nodejs/__test__/arrow.test.ts

@@ -1008,5 +1008,64 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
         expect(result).toEqual(null);
       });
     });
+
+    describe("boolean null handling", function () {
+      it("should handle null values in nullable boolean fields", () => {
+        const { makeArrowTable } = require("../lancedb/arrow");
+        const schema = new Schema([new Field("test", new arrow.Bool(), true)]);
+
+        // Test with all null values
+        const data = [{ test: null }];
+        const table = makeArrowTable(data, { schema });
+
+        expect(table.numRows).toBe(1);
+        expect(table.schema.names).toEqual(["test"]);
+        expect(table.getChild("test")!.get(0)).toBeNull();
+      });
+
+      it("should handle mixed null and non-null boolean values", () => {
+        const { makeArrowTable } = require("../lancedb/arrow");
+        const schema = new Schema([new Field("test", new Bool(), true)]);
+
+        // Test with mixed values
+        const data = [{ test: true }, { test: null }, { test: false }];
+        const table = makeArrowTable(data, { schema });
+
+        expect(table.numRows).toBe(3);
+        expect(table.getChild("test")!.get(0)).toBe(true);
+        expect(table.getChild("test")!.get(1)).toBeNull();
+        expect(table.getChild("test")!.get(2)).toBe(false);
+      });
+    });
   },
 );
+
+// Test for the undefined values bug fix
+describe("undefined values handling", () => {
+  it("should handle mixed undefined and actual values", () => {
+    const schema = new Schema([
+      new Field("text", new Utf8(), true), // nullable
+      new Field("number", new Int32(), true), // nullable
+      new Field("bool", new Bool(), true), // nullable
+    ]);
+
+    const data = [
+      { text: undefined, number: 42, bool: true },
+      { text: "hello", number: undefined, bool: false },
+      { text: "world", number: 123, bool: undefined },
+    ];
+    const table = makeArrowTable(data, { schema });
+
+    const result = table.toArray();
+    expect(result).toHaveLength(3);
+    expect(result[0].text).toBe(null);
+    expect(result[0].number).toBe(42);
+    expect(result[0].bool).toBe(true);
+    expect(result[1].text).toBe("hello");
+    expect(result[1].number).toBe(null);
+    expect(result[1].bool).toBe(false);
+    expect(result[2].text).toBe("world");
+    expect(result[2].number).toBe(123);
+    expect(result[2].bool).toBe(null);
+  });
+});

nodejs/__test__/connection.test.ts

@@ -203,3 +203,106 @@ describe("given a connection", () => {
     });
   });
 });
+
+describe("clone table functionality", () => {
+  let tmpDir: tmp.DirResult;
+  let db: Connection;
+  beforeEach(async () => {
+    tmpDir = tmp.dirSync({ unsafeCleanup: true });
+    db = await connect(tmpDir.name);
+  });
+  afterEach(() => tmpDir.removeCallback());
+
+  it("should clone a table with latest version (default behavior)", async () => {
+    // Create source table with some data
+    const data = [
+      { id: 1, text: "hello", vector: [1.0, 2.0] },
+      { id: 2, text: "world", vector: [3.0, 4.0] },
+    ];
+    const sourceTable = await db.createTable("source", data);
+
+    // Add more data to create a new version
+    const moreData = [{ id: 3, text: "test", vector: [5.0, 6.0] }];
+    await sourceTable.add(moreData);
+
+    // Clone the table (should get latest version with 3 rows)
+    const sourceUri = `${tmpDir.name}/source.lance`;
+    const clonedTable = await db.cloneTable("cloned", sourceUri);
+
+    // Verify cloned table has all 3 rows
+    expect(await clonedTable.countRows()).toBe(3);
+    expect((await db.tableNames()).includes("cloned")).toBe(true);
+  });
+
+  it("should clone a table from a specific version", async () => {
+    // Create source table with initial data
+    const data = [
+      { id: 1, text: "hello", vector: [1.0, 2.0] },
+      { id: 2, text: "world", vector: [3.0, 4.0] },
+    ];
+    const sourceTable = await db.createTable("source", data);
+
+    // Get the initial version
+    const initialVersion = await sourceTable.version();
+
+    // Add more data to create a new version
+    const moreData = [{ id: 3, text: "test", vector: [5.0, 6.0] }];
+    await sourceTable.add(moreData);
+
+    // Verify source now has 3 rows
+    expect(await sourceTable.countRows()).toBe(3);
+
+    // Clone from the initial version (should have only 2 rows)
+    const sourceUri = `${tmpDir.name}/source.lance`;
+    const clonedTable = await db.cloneTable("cloned", sourceUri, {
+      sourceVersion: initialVersion,
+    });
+
+    // Verify cloned table has only the initial 2 rows
+    expect(await clonedTable.countRows()).toBe(2);
+  });
+
+  it("should clone a table from a tagged version", async () => {
+    // Create source table with initial data
+    const data = [
+      { id: 1, text: "hello", vector: [1.0, 2.0] },
+      { id: 2, text: "world", vector: [3.0, 4.0] },
+    ];
+    const sourceTable = await db.createTable("source", data);
+
+    // Create a tag for the current version
+    const tags = await sourceTable.tags();
+    await tags.create("v1.0", await sourceTable.version());
+
+    // Add more data after the tag
+    const moreData = [{ id: 3, text: "test", vector: [5.0, 6.0] }];
+    await sourceTable.add(moreData);
+
+    // Verify source now has 3 rows
+    expect(await sourceTable.countRows()).toBe(3);
+
+    // Clone from the tagged version (should have only 2 rows)
+    const sourceUri = `${tmpDir.name}/source.lance`;
+    const clonedTable = await db.cloneTable("cloned", sourceUri, {
+      sourceTag: "v1.0",
+    });
+
+    // Verify cloned table has only the tagged version's 2 rows
+    expect(await clonedTable.countRows()).toBe(2);
+  });
+
+  it("should fail when attempting deep clone", async () => {
+    // Create source table with some data
+    const data = [
+      { id: 1, text: "hello", vector: [1.0, 2.0] },
+      { id: 2, text: "world", vector: [3.0, 4.0] },
+    ];
+    await db.createTable("source", data);
+
+    // Try to create a deep clone (should fail)
+    const sourceUri = `${tmpDir.name}/source.lance`;
+    await expect(
+      db.cloneTable("cloned", sourceUri, { isShallow: false }),
+    ).rejects.toThrow("Deep clone is not yet implemented");
+  });
+});

nodejs/__test__/embedding.test.ts

@@ -256,6 +256,60 @@ describe("embedding functions", () => {
     expect(actual).toHaveProperty("text");
   });
 
+  it("should handle undefined vector field with embedding function correctly", async () => {
+    @register("undefined_test")
+    class MockEmbeddingFunction extends EmbeddingFunction<string> {
+      ndims() {
+        return 3;
+      }
+      embeddingDataType(): Float {
+        return new Float32();
+      }
+      async computeQueryEmbeddings(_data: string) {
+        return [1, 2, 3];
+      }
+      async computeSourceEmbeddings(data: string[]) {
+        return Array.from({ length: data.length }).fill([
+          1, 2, 3,
+        ]) as number[][];
+      }
+    }
+    const func = getRegistry()
+      .get<MockEmbeddingFunction>("undefined_test")!
+      .create();
+    const schema = new Schema([
+      new Field("text", new Utf8(), true),
+      new Field(
+        "vector",
+        new FixedSizeList(3, new Field("item", new Float32(), true)),
+        true,
+      ),
+    ]);
+
+    const db = await connect(tmpDir.name);
+    const table = await db.createEmptyTable("test_undefined", schema, {
+      embeddingFunction: {
+        function: func,
+        sourceColumn: "text",
+        vectorColumn: "vector",
+      },
+    });
+
+    // Test that undefined, null, and omitted vector fields all work
+    await table.add([{ text: "test1", vector: undefined }]);
+    await table.add([{ text: "test2", vector: null }]);
+    await table.add([{ text: "test3" }]);
+
+    const rows = await table.query().toArray();
+    expect(rows.length).toBe(3);
+
+    // All rows should have vectors computed by the embedding function
+    for (const row of rows) {
+      expect(row.vector).toBeDefined();
+      expect(JSON.parse(JSON.stringify(row.vector))).toEqual([1, 2, 3]);
+    }
+  });
+
   test.each([new Float16(), new Float32(), new Float64()])(
     "should be able to provide manual embeddings with multiple float datatype",
     async (floatType) => {

nodejs/lancedb/arrow.ts

@@ -512,7 +512,11 @@ function* rowPathsAndValues(
     if (isObject(value)) {
       yield* rowPathsAndValues(value, [...basePath, key]);
     } else {
-      yield [[...basePath, key], value];
+      // Skip undefined values - they should be treated the same as missing fields
+      // for embedding function purposes
+      if (value !== undefined) {
+        yield [[...basePath, key], value];
+      }
     }
   }
 }
@@ -701,7 +705,7 @@ function transposeData(
       }
       return current;
     });
-    return makeVector(values, field.type);
+    return makeVector(values, field.type, undefined, field.nullable);
   }
 }
 
@@ -748,9 +752,30 @@ function makeVector(
   values: unknown[],
   type?: DataType,
   stringAsDictionary?: boolean,
+  nullable?: boolean,
   // biome-ignore lint/suspicious/noExplicitAny: skip
 ): Vector<any> {
   if (type !== undefined) {
+    // Convert undefined values to null for nullable fields
+    if (nullable) {
+      values = values.map((v) => (v === undefined ? null : v));
+    }
+
+    // workaround for: https://github.com/apache/arrow-js/issues/68
+    if (DataType.isBool(type)) {
+      const hasNonNullValue = values.some((v) => v !== null && v !== undefined);
+      if (!hasNonNullValue) {
+        const nullBitmap = new Uint8Array(Math.ceil(values.length / 8));
+        const data = makeData({
+          type: type,
+          length: values.length,
+          nullCount: values.length,
+          nullBitmap,
+        });
+        return arrowMakeVector(data);
+      }
+    }
+
     // No need for inference, let Arrow create it
     if (type instanceof Int) {
       if (DataType.isInt(type) && type.bitWidth === 64) {
@@ -875,7 +900,12 @@ async function applyEmbeddingsFromMetadata(
   for (const field of schema.fields) {
     if (!(field.name in columns)) {
       const nullValues = new Array(table.numRows).fill(null);
-      columns[field.name] = makeVector(nullValues, field.type);
+      columns[field.name] = makeVector(
+        nullValues,
+        field.type,
+        undefined,
+        field.nullable,
+      );
     }
   }
 
@@ -939,7 +969,12 @@ async function applyEmbeddings<T>(
     } else if (schema != null) {
       const destField = schema.fields.find((f) => f.name === destColumn);
       if (destField != null) {
-        newColumns[destColumn] = makeVector([], destField.type);
+        newColumns[destColumn] = makeVector(
+          [],
+          destField.type,
+          undefined,
+          destField.nullable,
+        );
       } else {
         throw new Error(
           `Attempt to apply embeddings to an empty table failed because schema was missing embedding column '${destColumn}'`,

nodejs/lancedb/connection.ts

@@ -268,6 +268,33 @@ export abstract class Connection {
    * @param {string[]} namespace The namespace to drop tables from (defaults to root namespace).
    */
   abstract dropAllTables(namespace?: string[]): Promise<void>;
+
+  /**
+   * Clone a table from a source table.
+   *
+   * A shallow clone creates a new table that shares the underlying data files
+   * with the source table but has its own independent manifest. This allows
+   * both the source and cloned tables to evolve independently while initially
+   * sharing the same data, deletion, and index files.
+   *
+   * @param {string} targetTableName - The name of the target table to create.
+   * @param {string} sourceUri - The URI of the source table to clone from.
+   * @param {object} options - Clone options.
+   * @param {string[]} options.targetNamespace - The namespace for the target table (defaults to root namespace).
+   * @param {number} options.sourceVersion - The version of the source table to clone.
+   * @param {string} options.sourceTag - The tag of the source table to clone.
+   * @param {boolean} options.isShallow - Whether to perform a shallow clone (defaults to true).
+   */
+  abstract cloneTable(
+    targetTableName: string,
+    sourceUri: string,
+    options?: {
+      targetNamespace?: string[];
+      sourceVersion?: number;
+      sourceTag?: string;
+      isShallow?: boolean;
+    },
+  ): Promise<Table>;
 }
 
 /** @hideconstructor */
@@ -332,6 +359,28 @@ export class LocalConnection extends Connection {
     return new LocalTable(innerTable);
   }
 
+  async cloneTable(
+    targetTableName: string,
+    sourceUri: string,
+    options?: {
+      targetNamespace?: string[];
+      sourceVersion?: number;
+      sourceTag?: string;
+      isShallow?: boolean;
+    },
+  ): Promise<Table> {
+    const innerTable = await this.inner.cloneTable(
+      targetTableName,
+      sourceUri,
+      options?.targetNamespace ?? [],
+      options?.sourceVersion ?? null,
+      options?.sourceTag ?? null,
+      options?.isShallow ?? true,
+    );
+
+    return new LocalTable(innerTable);
+  }
+
   private getStorageOptions(
     options?: Partial<CreateTableOptions>,
   ): Record<string, string> | undefined {

nodejs/npm/darwin-arm64/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-arm64",
-	"version": "0.22.1-beta.1",
+	"version": "0.22.1-beta.3",
 	"os": ["darwin"],
 	"cpu": ["arm64"],
 	"main": "lancedb.darwin-arm64.node",

nodejs/npm/darwin-x64/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-x64",
-	"version": "0.22.1-beta.1",
+	"version": "0.22.1-beta.3",
 	"os": ["darwin"],
 	"cpu": ["x64"],
 	"main": "lancedb.darwin-x64.node",

nodejs/npm/linux-arm64-gnu/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-gnu",
-	"version": "0.22.1-beta.1",
+	"version": "0.22.1-beta.3",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-gnu.node",

nodejs/npm/linux-arm64-musl/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-musl",
-	"version": "0.22.1-beta.1",
+	"version": "0.22.1-beta.3",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-musl.node",

nodejs/npm/linux-x64-gnu/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-gnu",
-	"version": "0.22.1-beta.1",
+	"version": "0.22.1-beta.3",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-gnu.node",

nodejs/npm/linux-x64-musl/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-musl",
-	"version": "0.22.1-beta.1",
+	"version": "0.22.1-beta.3",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-musl.node",

nodejs/npm/win32-arm64-msvc/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lancedb/lancedb-win32-arm64-msvc",
-  "version": "0.22.1-beta.1",
+  "version": "0.22.1-beta.3",
   "os": [
     "win32"
   ],

nodejs/npm/win32-x64-msvc/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-win32-x64-msvc",
-	"version": "0.22.1-beta.1",
+	"version": "0.22.1-beta.3",
 	"os": ["win32"],
 	"cpu": ["x64"],
 	"main": "lancedb.win32-x64-msvc.node",

nodejs/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "@lancedb/lancedb",
-  "version": "0.22.1-beta.1",
+  "version": "0.22.1-beta.3",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@lancedb/lancedb",
-      "version": "0.22.1-beta.1",
+      "version": "0.22.1-beta.3",
       "cpu": [
         "x64",
         "arm64"

nodejs/package.json

@@ -11,7 +11,7 @@
     "ann"
   ],
   "private": false,
-  "version": "0.22.1-beta.1",
+  "version": "0.22.1-beta.3",
   "main": "dist/index.js",
   "exports": {
     ".": "./dist/index.js",

nodejs/src/connection.rs

@@ -213,6 +213,36 @@ impl Connection {
         Ok(Table::new(tbl))
     }
 
+    #[napi(catch_unwind)]
+    pub async fn clone_table(
+        &self,
+        target_table_name: String,
+        source_uri: String,
+        target_namespace: Vec<String>,
+        source_version: Option<i64>,
+        source_tag: Option<String>,
+        is_shallow: bool,
+    ) -> napi::Result<Table> {
+        let mut builder = self
+            .get_inner()?
+            .clone_table(&target_table_name, &source_uri);
+
+        builder = builder.target_namespace(target_namespace);
+
+        if let Some(version) = source_version {
+            builder = builder.source_version(version as u64);
+        }
+
+        if let Some(tag) = source_tag {
+            builder = builder.source_tag(tag);
+        }
+
+        builder = builder.is_shallow(is_shallow);
+
+        let tbl = builder.execute().await.default_error()?;
+        Ok(Table::new(tbl))
+    }
+
     /// Drop table with the name. Or raise an error if the table does not exist.
     #[napi(catch_unwind)]
     pub async fn drop_table(&self, name: String, namespace: Vec<String>) -> napi::Result<()> {

python/.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.25.1-beta.2"
+current_version = "0.25.1"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

python/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "lancedb-python"
-version = "0.25.1-beta.2"
+version = "0.25.1"
 edition.workspace = true
 description = "Python bindings for LanceDB"
 license.workspace = true

python/python/lancedb/_lancedb.pyi

@@ -60,6 +60,15 @@ class Connection(object):
         storage_options: Optional[Dict[str, str]] = None,
         index_cache_size: Optional[int] = None,
     ) -> Table: ...
+    async def clone_table(
+        self,
+        target_table_name: str,
+        source_uri: str,
+        target_namespace: List[str] = [],
+        source_version: Optional[int] = None,
+        source_tag: Optional[str] = None,
+        is_shallow: bool = True,
+    ) -> Table: ...
     async def rename_table(
         self,
         cur_name: str,

python/python/lancedb/db.py

@@ -665,6 +665,60 @@ class LanceDBConnection(DBConnection):
             index_cache_size=index_cache_size,
         )
 
+    def clone_table(
+        self,
+        target_table_name: str,
+        source_uri: str,
+        *,
+        target_namespace: List[str] = [],
+        source_version: Optional[int] = None,
+        source_tag: Optional[str] = None,
+        is_shallow: bool = True,
+    ) -> LanceTable:
+        """Clone a table from a source table.
+
+        A shallow clone creates a new table that shares the underlying data files
+        with the source table but has its own independent manifest. This allows
+        both the source and cloned tables to evolve independently while initially
+        sharing the same data, deletion, and index files.
+
+        Parameters
+        ----------
+        target_table_name: str
+            The name of the target table to create.
+        source_uri: str
+            The URI of the source table to clone from.
+        target_namespace: List[str], optional
+            The namespace for the target table.
+            None or empty list represents root namespace.
+        source_version: int, optional
+            The version of the source table to clone.
+        source_tag: str, optional
+            The tag of the source table to clone.
+        is_shallow: bool, default True
+            Whether to perform a shallow clone (True) or deep clone (False).
+            Currently only shallow clone is supported.
+
+        Returns
+        -------
+        A LanceTable object representing the cloned table.
+        """
+        LOOP.run(
+            self._conn.clone_table(
+                target_table_name,
+                source_uri,
+                target_namespace=target_namespace,
+                source_version=source_version,
+                source_tag=source_tag,
+                is_shallow=is_shallow,
+            )
+        )
+        return LanceTable.open(
+            self,
+            target_table_name,
+            namespace=target_namespace,
+        )
+
     @override
     def drop_table(
         self,
@@ -1136,6 +1190,54 @@ class AsyncConnection(object):
         )
         return AsyncTable(table)
 
+    async def clone_table(
+        self,
+        target_table_name: str,
+        source_uri: str,
+        *,
+        target_namespace: List[str] = [],
+        source_version: Optional[int] = None,
+        source_tag: Optional[str] = None,
+        is_shallow: bool = True,
+    ) -> AsyncTable:
+        """Clone a table from a source table.
+
+        A shallow clone creates a new table that shares the underlying data files
+        with the source table but has its own independent manifest. This allows
+        both the source and cloned tables to evolve independently while initially
+        sharing the same data, deletion, and index files.
+
+        Parameters
+        ----------
+        target_table_name: str
+            The name of the target table to create.
+        source_uri: str
+            The URI of the source table to clone from.
+        target_namespace: List[str], optional
+            The namespace for the target table.
+            None or empty list represents root namespace.
+        source_version: int, optional
+            The version of the source table to clone.
+        source_tag: str, optional
+            The tag of the source table to clone.
+        is_shallow: bool, default True
+            Whether to perform a shallow clone (True) or deep clone (False).
+            Currently only shallow clone is supported.
+
+        Returns
+        -------
+        An AsyncTable object representing the cloned table.
+        """
+        table = await self._inner.clone_table(
+            target_table_name,
+            source_uri,
+            target_namespace=target_namespace,
+            source_version=source_version,
+            source_tag=source_tag,
+            is_shallow=is_shallow,
+        )
+        return AsyncTable(table)
+
     async def rename_table(
         self,
         cur_name: str,

python/python/lancedb/remote/db.py

@@ -212,6 +212,53 @@ class RemoteDBConnection(DBConnection):
         table = LOOP.run(self._conn.open_table(name, namespace=namespace))
         return RemoteTable(table, self.db_name)
 
+    def clone_table(
+        self,
+        target_table_name: str,
+        source_uri: str,
+        *,
+        target_namespace: List[str] = [],
+        source_version: Optional[int] = None,
+        source_tag: Optional[str] = None,
+        is_shallow: bool = True,
+    ) -> Table:
+        """Clone a table from a source table.
+
+        Parameters
+        ----------
+        target_table_name: str
+            The name of the target table to create.
+        source_uri: str
+            The URI of the source table to clone from.
+        target_namespace: List[str], optional
+            The namespace for the target table.
+            None or empty list represents root namespace.
+        source_version: int, optional
+            The version of the source table to clone.
+        source_tag: str, optional
+            The tag of the source table to clone.
+        is_shallow: bool, default True
+            Whether to perform a shallow clone (True) or deep clone (False).
+            Currently only shallow clone is supported.
+
+        Returns
+        -------
+        A RemoteTable object representing the cloned table.
+        """
+        from .table import RemoteTable
+
+        table = LOOP.run(
+            self._conn.clone_table(
+                target_table_name,
+                source_uri,
+                target_namespace=target_namespace,
+                source_version=source_version,
+                source_tag=source_tag,
+                is_shallow=is_shallow,
+            )
+        )
+        return RemoteTable(table, self.db_name)
+
     @override
     def create_table(
         self,

python/python/lancedb/rerankers/__init__.py

@@ -9,6 +9,7 @@ from .linear_combination import LinearCombinationReranker
 from .openai import OpenaiReranker
 from .jinaai import JinaReranker
 from .rrf import RRFReranker
+from .mrr import MRRReranker
 from .answerdotai import AnswerdotaiRerankers
 from .voyageai import VoyageAIReranker
 
@@ -23,4 +24,5 @@ __all__ = [
     "RRFReranker",
     "AnswerdotaiRerankers",
     "VoyageAIReranker",
+    "MRRReranker",
 ]

python/python/lancedb/rerankers/mrr.py

@@ -0,0 +1,169 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright The LanceDB Authors
+
+
+from typing import Union, List, TYPE_CHECKING
+import pyarrow as pa
+import numpy as np
+
+from collections import defaultdict
+from .base import Reranker
+
+if TYPE_CHECKING:
+    from ..table import LanceVectorQueryBuilder
+
+
+class MRRReranker(Reranker):
+    """
+    Reranks the results using Mean Reciprocal Rank (MRR) algorithm based
+    on the scores of vector and FTS search.
+    Algorithm reference - https://en.wikipedia.org/wiki/Mean_reciprocal_rank
+
+    MRR calculates the average of reciprocal ranks across different search results.
+    For each document, it computes the reciprocal of its rank in each system,
+    then takes the mean of these reciprocal ranks as the final score.
+
+    Parameters
+    ----------
+    weight_vector : float, default 0.5
+        Weight for vector search results (0.0 to 1.0)
+    weight_fts : float, default 0.5
+        Weight for FTS search results (0.0 to 1.0)
+        Note: weight_vector + weight_fts should equal 1.0
+    return_score : str, default "relevance"
+        Options are "relevance" or "all"
+        The type of score to return. If "relevance", will return only the relevance
+        score. If "all", will return all scores from the vector and FTS search along
+        with the relevance score.
+    """
+
+    def __init__(
+        self,
+        weight_vector: float = 0.5,
+        weight_fts: float = 0.5,
+        return_score="relevance",
+    ):
+        if not (0.0 <= weight_vector <= 1.0):
+            raise ValueError("weight_vector must be between 0.0 and 1.0")
+        if not (0.0 <= weight_fts <= 1.0):
+            raise ValueError("weight_fts must be between 0.0 and 1.0")
+        if abs(weight_vector + weight_fts - 1.0) > 1e-6:
+            raise ValueError("weight_vector + weight_fts must equal 1.0")
+
+        super().__init__(return_score)
+        self.weight_vector = weight_vector
+        self.weight_fts = weight_fts
+
+    def rerank_hybrid(
+        self,
+        query: str,  # noqa: F821
+        vector_results: pa.Table,
+        fts_results: pa.Table,
+    ):
+        vector_ids = vector_results["_rowid"].to_pylist() if vector_results else []
+        fts_ids = fts_results["_rowid"].to_pylist() if fts_results else []
+
+        # Maps result_id to list of (type, reciprocal_rank)
+        mrr_score_map = defaultdict(list)
+
+        if vector_ids:
+            for rank, result_id in enumerate(vector_ids, 1):
+                reciprocal_rank = 1.0 / rank
+                mrr_score_map[result_id].append(("vector", reciprocal_rank))
+
+        if fts_ids:
+            for rank, result_id in enumerate(fts_ids, 1):
+                reciprocal_rank = 1.0 / rank
+                mrr_score_map[result_id].append(("fts", reciprocal_rank))
+
+        final_mrr_scores = {}
+        for result_id, scores in mrr_score_map.items():
+            vector_rr = 0.0
+            fts_rr = 0.0
+
+            for score_type, reciprocal_rank in scores:
+                if score_type == "vector":
+                    vector_rr = reciprocal_rank
+                elif score_type == "fts":
+                    fts_rr = reciprocal_rank
+
+            # If a document doesn't appear, its reciprocal rank is 0
+            weighted_mrr = self.weight_vector * vector_rr + self.weight_fts * fts_rr
+            final_mrr_scores[result_id] = weighted_mrr
+
+        combined_results = self.merge_results(vector_results, fts_results)
+        combined_row_ids = combined_results["_rowid"].to_pylist()
+        relevance_scores = [final_mrr_scores[row_id] for row_id in combined_row_ids]
+        combined_results = combined_results.append_column(
+            "_relevance_score", pa.array(relevance_scores, type=pa.float32())
+        )
+        combined_results = combined_results.sort_by(
+            [("_relevance_score", "descending")]
+        )
+
+        if self.score == "relevance":
+            combined_results = self._keep_relevance_score(combined_results)
+
+        return combined_results
+
+    def rerank_multivector(
+        self,
+        vector_results: Union[List[pa.Table], List["LanceVectorQueryBuilder"]],
+        query: str = None,
+        deduplicate: bool = True,  # noqa: F821
+    ):
+        """
+        Reranks the results from multiple vector searches using MRR algorithm.
+        Each vector search result is treated as a separate ranking system,
+        and MRR calculates the mean of reciprocal ranks across all systems.
+        This cannot reuse rerank_hybrid because MRR semantics require treating
+        each vector result as a separate ranking system.
+        """
+        if not all(isinstance(v, type(vector_results[0])) for v in vector_results):
+            raise ValueError(
+                "All elements in vector_results should be of the same type"
+            )
+
+        # avoid circular import
+        if type(vector_results[0]).__name__ == "LanceVectorQueryBuilder":
+            vector_results = [result.to_arrow() for result in vector_results]
+        elif not isinstance(vector_results[0], pa.Table):
+            raise ValueError(
+                "vector_results should be a list of pa.Table or LanceVectorQueryBuilder"
+            )
+
+        if not all("_rowid" in result.column_names for result in vector_results):
+            raise ValueError(
+                "'_rowid' is required for deduplication. \
+                    add _rowid to search results like this: \
+                    `search().with_row_id(True)`"
+            )
+
+        mrr_score_map = defaultdict(list)
+
+        for result_table in vector_results:
+            result_ids = result_table["_rowid"].to_pylist()
+            for rank, result_id in enumerate(result_ids, 1):
+                reciprocal_rank = 1.0 / rank
+                mrr_score_map[result_id].append(reciprocal_rank)
+
+        final_mrr_scores = {}
+        for result_id, reciprocal_ranks in mrr_score_map.items():
+            mean_rr = np.mean(reciprocal_ranks)
+            final_mrr_scores[result_id] = mean_rr
+
+        combined = pa.concat_tables(vector_results, **self._concat_tables_args)
+        combined = self._deduplicate(combined)
+
+        combined_row_ids = combined["_rowid"].to_pylist()
+
+        relevance_scores = [final_mrr_scores[row_id] for row_id in combined_row_ids]
+        combined = combined.append_column(
+            "_relevance_score", pa.array(relevance_scores, type=pa.float32())
+        )
+        combined = combined.sort_by([("_relevance_score", "descending")])
+
+        if self.score == "relevance":
+            combined = self._keep_relevance_score(combined)
+
+        return combined

python/python/lancedb/table.py

@@ -1470,10 +1470,7 @@ class Table(ABC):
             be deleted unless they are at least 7 days old. If delete_unverified is True
             then these files will be deleted regardless of their age.
         retrain: bool, default False
-            If True, retrain the vector indices, this would refine the IVF clustering
+            This parameter is no longer used and is deprecated.
-            and quantization, which may improve the search accuracy. It's faster than
-            re-creating the index from scratch, so it's recommended to try this first,
-            when the data distribution has changed significantly.
 
         Experimental API
         ----------------
@@ -2835,10 +2832,7 @@ class LanceTable(Table):
             be deleted unless they are at least 7 days old. If delete_unverified is True
             then these files will be deleted regardless of their age.
         retrain: bool, default False
-            If True, retrain the vector indices, this would refine the IVF clustering
+            This parameter is no longer used and is deprecated.
-            and quantization, which may improve the search accuracy. It's faster than
-            re-creating the index from scratch, so it's recommended to try this first,
-            when the data distribution has changed significantly.
 
         Experimental API
         ----------------
@@ -4298,10 +4292,7 @@ class AsyncTable:
             be deleted unless they are at least 7 days old. If delete_unverified is True
             then these files will be deleted regardless of their age.
         retrain: bool, default False
-            If True, retrain the vector indices, this would refine the IVF clustering
+            This parameter is no longer used and is deprecated.
-            and quantization, which may improve the search accuracy. It's faster than
-            re-creating the index from scratch, so it's recommended to try this first,
-            when the data distribution has changed significantly.
 
         Experimental API
         ----------------
@@ -4324,10 +4315,19 @@ class AsyncTable:
         cleanup_since_ms: Optional[int] = None
         if cleanup_older_than is not None:
             cleanup_since_ms = round(cleanup_older_than.total_seconds() * 1000)
+
+        if retrain:
+            import warnings
+
+            warnings.warn(
+                "The 'retrain' parameter is deprecated and will be removed in a "
+                "future version.",
+                DeprecationWarning,
+            )
+
         return await self._inner.optimize(
             cleanup_since_ms=cleanup_since_ms,
             delete_unverified=delete_unverified,
-            retrain=retrain,
         )
 
     async def list_indices(self) -> Iterable[IndexConfig]:

python/python/tests/test_db.py

@@ -747,15 +747,16 @@ def test_local_namespace_operations(tmp_path):
     # Create a local database connection
     db = lancedb.connect(tmp_path)
 
-    # Test list_namespaces returns empty list
+    # Test list_namespaces returns empty list for root namespace
     namespaces = list(db.list_namespaces())
     assert namespaces == []
 
-    # Test list_namespaces with parameters still returns empty list
+    # Test list_namespaces with non-empty namespace raises NotImplementedError
-    namespaces_with_params = list(
+    with pytest.raises(
-        db.list_namespaces(namespace=["test"], page_token="token", limit=5)
+        NotImplementedError,
-    )
+        match="Namespace operations are not supported for listing database",
-    assert namespaces_with_params == []
+    ):
+        list(db.list_namespaces(namespace=["test"]))
 
 
 def test_local_create_namespace_not_supported(tmp_path):
@@ -830,3 +831,119 @@ def test_local_table_operations_with_namespace_raise_error(tmp_path):
     # Test table_names without namespace - should work normally
     tables_root = list(db.table_names())
     assert "test_table" in tables_root
+
+
+def test_clone_table_latest_version(tmp_path):
+    """Test cloning a table with the latest version (default behavior)"""
+    import os
+
+    db = lancedb.connect(tmp_path)
+
+    # Create source table with some data
+    data = [
+        {"id": 1, "text": "hello", "vector": [1.0, 2.0]},
+        {"id": 2, "text": "world", "vector": [3.0, 4.0]},
+    ]
+    source_table = db.create_table("source", data=data)
+
+    # Add more data to create a new version
+    more_data = [{"id": 3, "text": "test", "vector": [5.0, 6.0]}]
+    source_table.add(more_data)
+
+    # Clone the table (should get latest version with 3 rows)
+    source_uri = os.path.join(tmp_path, "source.lance")
+    cloned_table = db.clone_table("cloned", source_uri)
+
+    # Verify cloned table has all 3 rows
+    assert cloned_table.count_rows() == 3
+    assert "cloned" in db.table_names()
+
+    # Verify data matches
+    cloned_data = cloned_table.to_pandas()
+    assert len(cloned_data) == 3
+    assert set(cloned_data["id"].tolist()) == {1, 2, 3}
+
+
+def test_clone_table_specific_version(tmp_path):
+    """Test cloning a table from a specific version"""
+    import os
+
+    db = lancedb.connect(tmp_path)
+
+    # Create source table with initial data
+    data = [
+        {"id": 1, "text": "hello", "vector": [1.0, 2.0]},
+        {"id": 2, "text": "world", "vector": [3.0, 4.0]},
+    ]
+    source_table = db.create_table("source", data=data)
+
+    # Get the initial version
+    initial_version = source_table.version
+
+    # Add more data to create a new version
+    more_data = [{"id": 3, "text": "test", "vector": [5.0, 6.0]}]
+    source_table.add(more_data)
+
+    # Verify source now has 3 rows
+    assert source_table.count_rows() == 3
+
+    # Clone from the initial version (should have only 2 rows)
+    source_uri = os.path.join(tmp_path, "source.lance")
+    cloned_table = db.clone_table("cloned", source_uri, source_version=initial_version)
+
+    # Verify cloned table has only the initial 2 rows
+    assert cloned_table.count_rows() == 2
+    cloned_data = cloned_table.to_pandas()
+    assert set(cloned_data["id"].tolist()) == {1, 2}
+
+
+def test_clone_table_with_tag(tmp_path):
+    """Test cloning a table from a tagged version"""
+    import os
+
+    db = lancedb.connect(tmp_path)
+
+    # Create source table with initial data
+    data = [
+        {"id": 1, "text": "hello", "vector": [1.0, 2.0]},
+        {"id": 2, "text": "world", "vector": [3.0, 4.0]},
+    ]
+    source_table = db.create_table("source", data=data)
+
+    # Create a tag for the current version
+    source_table.tags.create("v1.0", source_table.version)
+
+    # Add more data after the tag
+    more_data = [{"id": 3, "text": "test", "vector": [5.0, 6.0]}]
+    source_table.add(more_data)
+
+    # Verify source now has 3 rows
+    assert source_table.count_rows() == 3
+
+    # Clone from the tagged version (should have only 2 rows)
+    source_uri = os.path.join(tmp_path, "source.lance")
+    cloned_table = db.clone_table("cloned", source_uri, source_tag="v1.0")
+
+    # Verify cloned table has only the tagged version's 2 rows
+    assert cloned_table.count_rows() == 2
+    cloned_data = cloned_table.to_pandas()
+    assert set(cloned_data["id"].tolist()) == {1, 2}
+
+
+def test_clone_table_deep_clone_fails(tmp_path):
+    """Test that deep clone raises an unsupported error"""
+    import os
+
+    db = lancedb.connect(tmp_path)
+
+    # Create source table with some data
+    data = [
+        {"id": 1, "text": "hello", "vector": [1.0, 2.0]},
+        {"id": 2, "text": "world", "vector": [3.0, 4.0]},
+    ]
+    db.create_table("source", data=data)
+
+    # Try to create a deep clone (should fail)
+    source_uri = os.path.join(tmp_path, "source.lance")
+    with pytest.raises(Exception, match="Deep clone is not yet implemented"):
+        db.clone_table("cloned", source_uri, is_shallow=False)

python/python/tests/test_rerankers.py

@@ -22,6 +22,7 @@ from lancedb.rerankers import (
     JinaReranker,
     AnswerdotaiRerankers,
     VoyageAIReranker,
+    MRRReranker,
 )
 from lancedb.table import LanceTable
 
@@ -46,6 +47,7 @@ def get_test_table(tmp_path, use_tantivy):
         db,
         "my_table",
         schema=MyTable,
+        mode="overwrite",
     )
 
     # Need to test with a bunch of phrases to make sure sorting is consistent
@@ -96,7 +98,7 @@ def get_test_table(tmp_path, use_tantivy):
     )
 
     # Create a fts index
-    table.create_fts_index("text", use_tantivy=use_tantivy)
+    table.create_fts_index("text", use_tantivy=use_tantivy, replace=True)
 
     return table, MyTable
 
@@ -320,6 +322,34 @@ def test_rrf_reranker(tmp_path, use_tantivy):
     _run_test_hybrid_reranker(reranker, tmp_path, use_tantivy)
 
 
+@pytest.mark.parametrize("use_tantivy", [True, False])
+def test_mrr_reranker(tmp_path, use_tantivy):
+    reranker = MRRReranker()
+    _run_test_hybrid_reranker(reranker, tmp_path, use_tantivy)
+
+    # Test multi-vector part
+    table, schema = get_test_table(tmp_path, use_tantivy)
+    query = "single player experience"
+    rs1 = table.search(query, vector_column_name="vector").limit(10).with_row_id(True)
+    rs2 = (
+        table.search(query, vector_column_name="meta_vector")
+        .limit(10)
+        .with_row_id(True)
+    )
+    result = reranker.rerank_multivector([rs1, rs2])
+    assert "_relevance_score" in result.column_names
+    assert len(result) <= 20
+
+    if len(result) > 1:
+        assert np.all(np.diff(result.column("_relevance_score").to_numpy()) <= 0), (
+            "The _relevance_score should be descending."
+        )
+
+    # Test with duplicate results
+    result_deduped = reranker.rerank_multivector([rs1, rs2, rs1])
+    assert len(result_deduped) == len(result)
+
+
 def test_rrf_reranker_distance():
     data = pa.table(
         {

python/src/connection.rs

@@ -163,6 +163,34 @@ impl Connection {
         })
     }
 
+    #[pyo3(signature = (target_table_name, source_uri, target_namespace=vec![], source_version=None, source_tag=None, is_shallow=true))]
+    pub fn clone_table(
+        self_: PyRef<'_, Self>,
+        target_table_name: String,
+        source_uri: String,
+        target_namespace: Vec<String>,
+        source_version: Option<u64>,
+        source_tag: Option<String>,
+        is_shallow: bool,
+    ) -> PyResult<Bound<'_, PyAny>> {
+        let inner = self_.get_inner()?.clone();
+
+        let mut builder = inner.clone_table(target_table_name, source_uri);
+        builder = builder.target_namespace(target_namespace);
+        if let Some(version) = source_version {
+            builder = builder.source_version(version);
+        }
+        if let Some(tag) = source_tag {
+            builder = builder.source_tag(tag);
+        }
+        builder = builder.is_shallow(is_shallow);
+
+        future_into_py(self_.py(), async move {
+            let table = builder.execute().await.infer_error()?;
+            Ok(Table::new(table))
+        })
+    }
+
     #[pyo3(signature = (cur_name, new_name, cur_namespace=vec![], new_namespace=vec![]))]
     pub fn rename_table(
         self_: PyRef<'_, Self>,

python/src/table.rs

@@ -591,12 +591,11 @@ impl Table {
     }
 
     /// Optimize the on-disk data by compacting and pruning old data, for better performance.
-    #[pyo3(signature = (cleanup_since_ms=None, delete_unverified=None, retrain=None))]
+    #[pyo3(signature = (cleanup_since_ms=None, delete_unverified=None))]
     pub fn optimize(
         self_: PyRef<'_, Self>,
         cleanup_since_ms: Option<u64>,
         delete_unverified: Option<bool>,
-        retrain: Option<bool>,
     ) -> PyResult<Bound<'_, PyAny>> {
         let inner = self_.inner_ref()?.clone();
         let older_than = if let Some(ms) = cleanup_since_ms {
@@ -632,10 +631,9 @@ impl Table {
                 .prune
                 .unwrap();
             inner
-                .optimize(lancedb::table::OptimizeAction::Index(match retrain {
+                .optimize(lancedb::table::OptimizeAction::Index(
-                    Some(true) => OptimizeOptions::retrain(),
+                    OptimizeOptions::default(),
-                    _ => OptimizeOptions::default(),
+                ))
-                }))
                 .await
                 .infer_error()?;
             Ok(OptimizeStats {

rust/lancedb/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "lancedb"
-version = "0.22.1-beta.1"
+version = "0.22.1-beta.3"
 edition.workspace = true
 description = "LanceDB: A serverless, low-latency vector database for AI applications"
 license.workspace = true

rust/lancedb/src/connection.rs

@@ -17,9 +17,9 @@ use crate::database::listing::{
     ListingDatabase, OPT_NEW_TABLE_STORAGE_VERSION, OPT_NEW_TABLE_V2_MANIFEST_PATHS,
 };
 use crate::database::{
-    CreateNamespaceRequest, CreateTableData, CreateTableMode, CreateTableRequest, Database,
+    CloneTableRequest, CreateNamespaceRequest, CreateTableData, CreateTableMode,
-    DatabaseOptions, DropNamespaceRequest, ListNamespacesRequest, OpenTableRequest,
+    CreateTableRequest, Database, DatabaseOptions, DropNamespaceRequest, ListNamespacesRequest,
-    TableNamesRequest,
+    OpenTableRequest, TableNamesRequest,
 };
 use crate::embeddings::{
     EmbeddingDefinition, EmbeddingFunction, EmbeddingRegistry, MemoryRegistry, WithEmbeddings,
@@ -469,6 +469,62 @@ impl OpenTableBuilder {
     }
 }
 
+/// Builder for cloning a table.
+///
+/// A shallow clone creates a new table that shares the underlying data files
+/// with the source table but has its own independent manifest. Both the source
+/// and cloned tables can evolve independently while initially sharing the same
+/// data, deletion, and index files.
+///
+/// Use this builder to configure the clone operation before executing it.
+pub struct CloneTableBuilder {
+    parent: Arc<dyn Database>,
+    request: CloneTableRequest,
+}
+
+impl CloneTableBuilder {
+    fn new(parent: Arc<dyn Database>, target_table_name: String, source_uri: String) -> Self {
+        Self {
+            parent,
+            request: CloneTableRequest::new(target_table_name, source_uri),
+        }
+    }
+
+    /// Set the source version to clone from
+    pub fn source_version(mut self, version: u64) -> Self {
+        self.request.source_version = Some(version);
+        self
+    }
+
+    /// Set the source tag to clone from
+    pub fn source_tag(mut self, tag: impl Into<String>) -> Self {
+        self.request.source_tag = Some(tag.into());
+        self
+    }
+
+    /// Set the target namespace for the cloned table
+    pub fn target_namespace(mut self, namespace: Vec<String>) -> Self {
+        self.request.target_namespace = namespace;
+        self
+    }
+
+    /// Set whether to perform a shallow clone (default: true)
+    ///
+    /// When true, the cloned table shares data files with the source table.
+    /// When false, performs a deep clone (not yet implemented).
+    pub fn is_shallow(mut self, is_shallow: bool) -> Self {
+        self.request.is_shallow = is_shallow;
+        self
+    }
+
+    /// Execute the clone operation
+    pub async fn execute(self) -> Result<Table> {
+        Ok(Table::new(
+            self.parent.clone().clone_table(self.request).await?,
+        ))
+    }
+}
+
 /// A connection to LanceDB
 #[derive(Clone)]
 pub struct Connection {
@@ -575,6 +631,30 @@ impl Connection {
         )
     }
 
+    /// Clone a table in the database
+    ///
+    /// Creates a new table by cloning from an existing source table.
+    /// By default, this performs a shallow clone where the new table shares
+    /// the underlying data files with the source table.
+    ///
+    /// # Parameters
+    /// - `target_table_name`: The name of the new table to create
+    /// - `source_uri`: The URI of the source table to clone from
+    ///
+    /// # Returns
+    /// A [`CloneTableBuilder`] that can be used to configure the clone operation
+    pub fn clone_table(
+        &self,
+        target_table_name: impl Into<String>,
+        source_uri: impl Into<String>,
+    ) -> CloneTableBuilder {
+        CloneTableBuilder::new(
+            self.internal.clone(),
+            target_table_name.into(),
+            source_uri.into(),
+        )
+    }
+
     /// Rename a table in the database.
     ///
     /// This is only supported in LanceDB Cloud.
@@ -1281,4 +1361,50 @@ mod tests {
             .unwrap();
         assert_eq!(other_schema, overwritten.schema().await.unwrap());
     }
+
+    #[tokio::test]
+    async fn test_clone_table() {
+        let tmp_dir = tempdir().unwrap();
+        let uri = tmp_dir.path().to_str().unwrap();
+        let db = connect(uri).execute().await.unwrap();
+
+        // Create a source table with some data
+        let mut batch_gen = BatchGenerator::new()
+            .col(Box::new(IncrementingInt32::new().named("id")))
+            .col(Box::new(IncrementingInt32::new().named("value")));
+        let reader = batch_gen.batches(5, 100);
+
+        let source_table = db
+            .create_table("source_table", reader)
+            .execute()
+            .await
+            .unwrap();
+
+        // Get the source table URI
+        let source_table_path = tmp_dir.path().join("source_table.lance");
+        let source_uri = source_table_path.to_str().unwrap();
+
+        // Clone the table
+        let cloned_table = db
+            .clone_table("cloned_table", source_uri)
+            .execute()
+            .await
+            .unwrap();
+
+        // Verify the cloned table exists
+        let table_names = db.table_names().execute().await.unwrap();
+        assert!(table_names.contains(&"source_table".to_string()));
+        assert!(table_names.contains(&"cloned_table".to_string()));
+
+        // Verify the cloned table has the same schema
+        assert_eq!(
+            source_table.schema().await.unwrap(),
+            cloned_table.schema().await.unwrap()
+        );
+
+        // Verify the cloned table has the same data
+        let source_count = source_table.count_rows(None).await.unwrap();
+        let cloned_count = cloned_table.count_rows(None).await.unwrap();
+        assert_eq!(source_count, cloned_count);
+    }
 }

rust/lancedb/src/database.rs

@@ -176,6 +176,42 @@ impl CreateTableRequest {
     }
 }
 
+/// Request to clone a table from a source table.
+///
+/// A shallow clone creates a new table that shares the underlying data files
+/// with the source table but has its own independent manifest. This allows
+/// both the source and cloned tables to evolve independently while initially
+/// sharing the same data, deletion, and index files.
+#[derive(Clone, Debug)]
+pub struct CloneTableRequest {
+    /// The name of the target table to create
+    pub target_table_name: String,
+    /// The namespace for the target table. Empty list represents root namespace.
+    pub target_namespace: Vec<String>,
+    /// The URI of the source table to clone from.
+    pub source_uri: String,
+    /// Optional version of the source table to clone.
+    pub source_version: Option<u64>,
+    /// Optional tag of the source table to clone.
+    pub source_tag: Option<String>,
+    /// Whether to perform a shallow clone (true) or deep clone (false). Defaults to true.
+    /// Currently only shallow clone is supported.
+    pub is_shallow: bool,
+}
+
+impl CloneTableRequest {
+    pub fn new(target_table_name: String, source_uri: String) -> Self {
+        Self {
+            target_table_name,
+            target_namespace: vec![],
+            source_uri,
+            source_version: None,
+            source_tag: None,
+            is_shallow: true,
+        }
+    }
+}
+
 /// The `Database` trait defines the interface for database implementations.
 ///
 /// A database is responsible for managing tables and their metadata.
@@ -193,6 +229,13 @@ pub trait Database:
     async fn table_names(&self, request: TableNamesRequest) -> Result<Vec<String>>;
     /// Create a table in the database
     async fn create_table(&self, request: CreateTableRequest) -> Result<Arc<dyn BaseTable>>;
+    /// Clone a table in the database.
+    ///
+    /// Creates a shallow clone of the source table, sharing underlying data files
+    /// but with an independent manifest. Both tables can evolve separately after cloning.
+    ///
+    /// See [`CloneTableRequest`] for detailed documentation and examples.
+    async fn clone_table(&self, request: CloneTableRequest) -> Result<Arc<dyn BaseTable>>;
     /// Open a table in the database
     async fn open_table(&self, request: OpenTableRequest) -> Result<Arc<dyn BaseTable>>;
     /// Rename a table in the database

rust/lancedb/src/database/listing.rs

@@ -7,7 +7,8 @@ use std::fs::create_dir_all;
 use std::path::Path;
 use std::{collections::HashMap, sync::Arc};
 
-use lance::dataset::{ReadParams, WriteMode};
+use lance::dataset::refs::Ref;
+use lance::dataset::{builder::DatasetBuilder, ReadParams, WriteMode};
 use lance::io::{ObjectStore, ObjectStoreParams, WrappingObjectStore};
 use lance_datafusion::utils::StreamingWriteSource;
 use lance_encoding::version::LanceFileVersion;
@@ -22,8 +23,8 @@ use crate::table::NativeTable;
 use crate::utils::validate_table_name;
 
 use super::{
-    BaseTable, CreateNamespaceRequest, CreateTableMode, CreateTableRequest, Database,
+    BaseTable, CloneTableRequest, CreateNamespaceRequest, CreateTableMode, CreateTableRequest,
-    DatabaseOptions, DropNamespaceRequest, ListNamespacesRequest, OpenTableRequest,
+    Database, DatabaseOptions, DropNamespaceRequest, ListNamespacesRequest, OpenTableRequest,
     TableNamesRequest,
 };
 
@@ -684,6 +685,65 @@ impl Database for ListingDatabase {
         }
     }
 
+    async fn clone_table(&self, request: CloneTableRequest) -> Result<Arc<dyn BaseTable>> {
+        if !request.target_namespace.is_empty() {
+            return Err(Error::NotSupported {
+                message: "Namespace parameter is not supported for listing database. Only root namespace is supported.".into(),
+            });
+        }
+
+        // TODO: support deep clone
+        if !request.is_shallow {
+            return Err(Error::NotSupported {
+                message: "Deep clone is not yet implemented".to_string(),
+            });
+        }
+
+        validate_table_name(&request.target_table_name)?;
+
+        let storage_params = ObjectStoreParams {
+            storage_options: Some(self.storage_options.clone()),
+            ..Default::default()
+        };
+        let read_params = ReadParams {
+            store_options: Some(storage_params.clone()),
+            session: Some(self.session.clone()),
+            ..Default::default()
+        };
+
+        let mut source_dataset = DatasetBuilder::from_uri(&request.source_uri)
+            .with_read_params(read_params.clone())
+            .load()
+            .await
+            .map_err(|e| Error::Lance { source: e })?;
+
+        let version_ref = match (request.source_version, request.source_tag) {
+            (Some(v), None) => Ok(Ref::Version(v)),
+            (None, Some(tag)) => Ok(Ref::Tag(tag)),
+            (None, None) => Ok(Ref::Version(source_dataset.version().version)),
+            _ => Err(Error::InvalidInput {
+                message: "Cannot specify both source_version and source_tag".to_string(),
+            }),
+        }?;
+
+        let target_uri = self.table_uri(&request.target_table_name)?;
+        source_dataset
+            .shallow_clone(&target_uri, version_ref, storage_params)
+            .await
+            .map_err(|e| Error::Lance { source: e })?;
+
+        let cloned_table = NativeTable::open_with_params(
+            &target_uri,
+            &request.target_table_name,
+            self.store_wrapper.clone(),
+            None,
+            self.read_consistency_interval,
+        )
+        .await?;
+
+        Ok(Arc::new(cloned_table))
+    }
+
     async fn open_table(&self, mut request: OpenTableRequest) -> Result<Arc<dyn BaseTable>> {
         if !request.namespace.is_empty() {
             return Err(Error::NotSupported {
@@ -785,3 +845,694 @@ impl Database for ListingDatabase {
         self
     }
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::connection::ConnectRequest;
+    use crate::database::{CreateTableData, CreateTableMode, CreateTableRequest};
+    use crate::table::{Table, TableDefinition};
+    use arrow_array::{Int32Array, RecordBatch, StringArray};
+    use arrow_schema::{DataType, Field, Schema};
+    use tempfile::tempdir;
+
+    async fn setup_database() -> (tempfile::TempDir, ListingDatabase) {
+        let tempdir = tempdir().unwrap();
+        let uri = tempdir.path().to_str().unwrap();
+
+        let request = ConnectRequest {
+            uri: uri.to_string(),
+            #[cfg(feature = "remote")]
+            client_config: Default::default(),
+            options: Default::default(),
+            read_consistency_interval: None,
+            session: None,
+        };
+
+        let db = ListingDatabase::connect_with_options(&request)
+            .await
+            .unwrap();
+
+        (tempdir, db)
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_basic() {
+        let (_tempdir, db) = setup_database().await;
+
+        // Create a source table with schema
+        let schema = Arc::new(Schema::new(vec![
+            Field::new("id", DataType::Int32, false),
+            Field::new("name", DataType::Utf8, false),
+        ]));
+
+        let source_table = db
+            .create_table(CreateTableRequest {
+                name: "source_table".to_string(),
+                namespace: vec![],
+                data: CreateTableData::Empty(TableDefinition::new_from_schema(schema.clone())),
+                mode: CreateTableMode::Create,
+                write_options: Default::default(),
+            })
+            .await
+            .unwrap();
+
+        // Get the source table URI
+        let source_uri = db.table_uri("source_table").unwrap();
+
+        // Clone the table
+        let cloned_table = db
+            .clone_table(CloneTableRequest {
+                target_table_name: "cloned_table".to_string(),
+                target_namespace: vec![],
+                source_uri: source_uri.clone(),
+                source_version: None,
+                source_tag: None,
+                is_shallow: true,
+            })
+            .await
+            .unwrap();
+
+        // Verify both tables exist
+        let table_names = db.table_names(TableNamesRequest::default()).await.unwrap();
+        assert!(table_names.contains(&"source_table".to_string()));
+        assert!(table_names.contains(&"cloned_table".to_string()));
+
+        // Verify schemas match
+        assert_eq!(
+            source_table.schema().await.unwrap(),
+            cloned_table.schema().await.unwrap()
+        );
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_with_data() {
+        let (_tempdir, db) = setup_database().await;
+
+        // Create a source table with actual data
+        let schema = Arc::new(Schema::new(vec![
+            Field::new("id", DataType::Int32, false),
+            Field::new("name", DataType::Utf8, false),
+        ]));
+
+        let batch = RecordBatch::try_new(
+            schema.clone(),
+            vec![
+                Arc::new(Int32Array::from(vec![1, 2, 3])),
+                Arc::new(StringArray::from(vec!["a", "b", "c"])),
+            ],
+        )
+        .unwrap();
+
+        let reader = Box::new(arrow_array::RecordBatchIterator::new(
+            vec![Ok(batch)],
+            schema.clone(),
+        ));
+
+        let source_table = db
+            .create_table(CreateTableRequest {
+                name: "source_with_data".to_string(),
+                namespace: vec![],
+                data: CreateTableData::Data(reader),
+                mode: CreateTableMode::Create,
+                write_options: Default::default(),
+            })
+            .await
+            .unwrap();
+
+        let source_uri = db.table_uri("source_with_data").unwrap();
+
+        // Clone the table
+        let cloned_table = db
+            .clone_table(CloneTableRequest {
+                target_table_name: "cloned_with_data".to_string(),
+                target_namespace: vec![],
+                source_uri,
+                source_version: None,
+                source_tag: None,
+                is_shallow: true,
+            })
+            .await
+            .unwrap();
+
+        // Verify data counts match
+        let source_count = source_table.count_rows(None).await.unwrap();
+        let cloned_count = cloned_table.count_rows(None).await.unwrap();
+        assert_eq!(source_count, cloned_count);
+        assert_eq!(source_count, 3);
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_with_storage_options() {
+        let tempdir = tempdir().unwrap();
+        let uri = tempdir.path().to_str().unwrap();
+
+        // Create database with storage options
+        let mut options = HashMap::new();
+        options.insert("test_option".to_string(), "test_value".to_string());
+
+        let request = ConnectRequest {
+            uri: uri.to_string(),
+            #[cfg(feature = "remote")]
+            client_config: Default::default(),
+            options: options.clone(),
+            read_consistency_interval: None,
+            session: None,
+        };
+
+        let db = ListingDatabase::connect_with_options(&request)
+            .await
+            .unwrap();
+
+        // Create source table
+        let schema = Arc::new(Schema::new(vec![Field::new("id", DataType::Int32, false)]));
+
+        db.create_table(CreateTableRequest {
+            name: "source".to_string(),
+            namespace: vec![],
+            data: CreateTableData::Empty(TableDefinition::new_from_schema(schema)),
+            mode: CreateTableMode::Create,
+            write_options: Default::default(),
+        })
+        .await
+        .unwrap();
+
+        let source_uri = db.table_uri("source").unwrap();
+
+        // Clone should work with storage options
+        let cloned = db
+            .clone_table(CloneTableRequest {
+                target_table_name: "cloned".to_string(),
+                target_namespace: vec![],
+                source_uri,
+                source_version: None,
+                source_tag: None,
+                is_shallow: true,
+            })
+            .await;
+
+        assert!(cloned.is_ok());
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_deep_not_supported() {
+        let (_tempdir, db) = setup_database().await;
+
+        // Create a source table
+        let schema = Arc::new(Schema::new(vec![Field::new("id", DataType::Int32, false)]));
+
+        db.create_table(CreateTableRequest {
+            name: "source".to_string(),
+            namespace: vec![],
+            data: CreateTableData::Empty(TableDefinition::new_from_schema(schema)),
+            mode: CreateTableMode::Create,
+            write_options: Default::default(),
+        })
+        .await
+        .unwrap();
+
+        let source_uri = db.table_uri("source").unwrap();
+
+        // Try deep clone (should fail)
+        let result = db
+            .clone_table(CloneTableRequest {
+                target_table_name: "cloned".to_string(),
+                target_namespace: vec![],
+                source_uri,
+                source_version: None,
+                source_tag: None,
+                is_shallow: false, // Request deep clone
+            })
+            .await;
+
+        assert!(result.is_err());
+        assert!(matches!(
+            result.unwrap_err(),
+            Error::NotSupported { message } if message.contains("Deep clone")
+        ));
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_with_namespace_not_supported() {
+        let (_tempdir, db) = setup_database().await;
+
+        // Create a source table
+        let schema = Arc::new(Schema::new(vec![Field::new("id", DataType::Int32, false)]));
+
+        db.create_table(CreateTableRequest {
+            name: "source".to_string(),
+            namespace: vec![],
+            data: CreateTableData::Empty(TableDefinition::new_from_schema(schema)),
+            mode: CreateTableMode::Create,
+            write_options: Default::default(),
+        })
+        .await
+        .unwrap();
+
+        let source_uri = db.table_uri("source").unwrap();
+
+        // Try clone with namespace (should fail for listing database)
+        let result = db
+            .clone_table(CloneTableRequest {
+                target_table_name: "cloned".to_string(),
+                target_namespace: vec!["namespace".to_string()], // Non-empty namespace
+                source_uri,
+                source_version: None,
+                source_tag: None,
+                is_shallow: true,
+            })
+            .await;
+
+        assert!(result.is_err());
+        assert!(matches!(
+            result.unwrap_err(),
+            Error::NotSupported { message } if message.contains("Namespace parameter is not supported")
+        ));
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_invalid_target_name() {
+        let (_tempdir, db) = setup_database().await;
+
+        // Create a source table
+        let schema = Arc::new(Schema::new(vec![Field::new("id", DataType::Int32, false)]));
+
+        db.create_table(CreateTableRequest {
+            name: "source".to_string(),
+            namespace: vec![],
+            data: CreateTableData::Empty(TableDefinition::new_from_schema(schema)),
+            mode: CreateTableMode::Create,
+            write_options: Default::default(),
+        })
+        .await
+        .unwrap();
+
+        let source_uri = db.table_uri("source").unwrap();
+
+        // Try clone with invalid target name
+        let result = db
+            .clone_table(CloneTableRequest {
+                target_table_name: "invalid/name".to_string(), // Invalid name with slash
+                target_namespace: vec![],
+                source_uri,
+                source_version: None,
+                source_tag: None,
+                is_shallow: true,
+            })
+            .await;
+
+        assert!(result.is_err());
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_source_not_found() {
+        let (_tempdir, db) = setup_database().await;
+
+        // Try to clone from non-existent source
+        let result = db
+            .clone_table(CloneTableRequest {
+                target_table_name: "cloned".to_string(),
+                target_namespace: vec![],
+                source_uri: "/nonexistent/table.lance".to_string(),
+                source_version: None,
+                source_tag: None,
+                is_shallow: true,
+            })
+            .await;
+
+        assert!(result.is_err());
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_with_version_and_tag_error() {
+        let (_tempdir, db) = setup_database().await;
+
+        // Create a source table
+        let schema = Arc::new(Schema::new(vec![Field::new("id", DataType::Int32, false)]));
+
+        db.create_table(CreateTableRequest {
+            name: "source".to_string(),
+            namespace: vec![],
+            data: CreateTableData::Empty(TableDefinition::new_from_schema(schema)),
+            mode: CreateTableMode::Create,
+            write_options: Default::default(),
+        })
+        .await
+        .unwrap();
+
+        let source_uri = db.table_uri("source").unwrap();
+
+        // Try clone with both version and tag (should fail)
+        let result = db
+            .clone_table(CloneTableRequest {
+                target_table_name: "cloned".to_string(),
+                target_namespace: vec![],
+                source_uri,
+                source_version: Some(1),
+                source_tag: Some("v1.0".to_string()),
+                is_shallow: true,
+            })
+            .await;
+
+        assert!(result.is_err());
+        assert!(matches!(
+            result.unwrap_err(),
+            Error::InvalidInput { message } if message.contains("Cannot specify both source_version and source_tag")
+        ));
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_with_specific_version() {
+        let (_tempdir, db) = setup_database().await;
+
+        // Create a source table with initial data
+        let schema = Arc::new(Schema::new(vec![
+            Field::new("id", DataType::Int32, false),
+            Field::new("value", DataType::Utf8, false),
+        ]));
+
+        let batch1 = RecordBatch::try_new(
+            schema.clone(),
+            vec![
+                Arc::new(Int32Array::from(vec![1, 2])),
+                Arc::new(StringArray::from(vec!["a", "b"])),
+            ],
+        )
+        .unwrap();
+
+        let reader = Box::new(arrow_array::RecordBatchIterator::new(
+            vec![Ok(batch1)],
+            schema.clone(),
+        ));
+
+        let source_table = db
+            .create_table(CreateTableRequest {
+                name: "versioned_source".to_string(),
+                namespace: vec![],
+                data: CreateTableData::Data(reader),
+                mode: CreateTableMode::Create,
+                write_options: Default::default(),
+            })
+            .await
+            .unwrap();
+
+        // Get the initial version
+        let initial_version = source_table.version().await.unwrap();
+
+        // Add more data to create a new version
+        let batch2 = RecordBatch::try_new(
+            schema.clone(),
+            vec![
+                Arc::new(Int32Array::from(vec![3, 4])),
+                Arc::new(StringArray::from(vec!["c", "d"])),
+            ],
+        )
+        .unwrap();
+
+        let source_table_obj = Table::new(source_table.clone());
+        source_table_obj
+            .add(Box::new(arrow_array::RecordBatchIterator::new(
+                vec![Ok(batch2)],
+                schema.clone(),
+            )))
+            .execute()
+            .await
+            .unwrap();
+
+        // Verify source table now has 4 rows
+        assert_eq!(source_table.count_rows(None).await.unwrap(), 4);
+
+        let source_uri = db.table_uri("versioned_source").unwrap();
+
+        // Clone from the initial version (should have only 2 rows)
+        let cloned_table = db
+            .clone_table(CloneTableRequest {
+                target_table_name: "cloned_from_version".to_string(),
+                target_namespace: vec![],
+                source_uri,
+                source_version: Some(initial_version),
+                source_tag: None,
+                is_shallow: true,
+            })
+            .await
+            .unwrap();
+
+        // Verify cloned table has only the initial 2 rows
+        assert_eq!(cloned_table.count_rows(None).await.unwrap(), 2);
+
+        // Source table should still have 4 rows
+        assert_eq!(source_table.count_rows(None).await.unwrap(), 4);
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_with_tag() {
+        let (_tempdir, db) = setup_database().await;
+
+        // Create a source table with initial data
+        let schema = Arc::new(Schema::new(vec![
+            Field::new("id", DataType::Int32, false),
+            Field::new("value", DataType::Utf8, false),
+        ]));
+
+        let batch1 = RecordBatch::try_new(
+            schema.clone(),
+            vec![
+                Arc::new(Int32Array::from(vec![1, 2])),
+                Arc::new(StringArray::from(vec!["a", "b"])),
+            ],
+        )
+        .unwrap();
+
+        let reader = Box::new(arrow_array::RecordBatchIterator::new(
+            vec![Ok(batch1)],
+            schema.clone(),
+        ));
+
+        let source_table = db
+            .create_table(CreateTableRequest {
+                name: "tagged_source".to_string(),
+                namespace: vec![],
+                data: CreateTableData::Data(reader),
+                mode: CreateTableMode::Create,
+                write_options: Default::default(),
+            })
+            .await
+            .unwrap();
+
+        // Create a tag for the current version
+        let source_table_obj = Table::new(source_table.clone());
+        let mut tags = source_table_obj.tags().await.unwrap();
+        tags.create("v1.0", source_table.version().await.unwrap())
+            .await
+            .unwrap();
+
+        // Add more data after the tag
+        let batch2 = RecordBatch::try_new(
+            schema.clone(),
+            vec![
+                Arc::new(Int32Array::from(vec![3, 4])),
+                Arc::new(StringArray::from(vec!["c", "d"])),
+            ],
+        )
+        .unwrap();
+
+        let source_table_obj = Table::new(source_table.clone());
+        source_table_obj
+            .add(Box::new(arrow_array::RecordBatchIterator::new(
+                vec![Ok(batch2)],
+                schema.clone(),
+            )))
+            .execute()
+            .await
+            .unwrap();
+
+        // Source table should have 4 rows
+        assert_eq!(source_table.count_rows(None).await.unwrap(), 4);
+
+        let source_uri = db.table_uri("tagged_source").unwrap();
+
+        // Clone from the tag (should have only 2 rows)
+        let cloned_table = db
+            .clone_table(CloneTableRequest {
+                target_table_name: "cloned_from_tag".to_string(),
+                target_namespace: vec![],
+                source_uri,
+                source_version: None,
+                source_tag: Some("v1.0".to_string()),
+                is_shallow: true,
+            })
+            .await
+            .unwrap();
+
+        // Verify cloned table has only the tagged version's 2 rows
+        assert_eq!(cloned_table.count_rows(None).await.unwrap(), 2);
+    }
+
+    #[tokio::test]
+    async fn test_cloned_tables_evolve_independently() {
+        let (_tempdir, db) = setup_database().await;
+
+        // Create a source table with initial data
+        let schema = Arc::new(Schema::new(vec![
+            Field::new("id", DataType::Int32, false),
+            Field::new("value", DataType::Utf8, false),
+        ]));
+
+        let batch1 = RecordBatch::try_new(
+            schema.clone(),
+            vec![
+                Arc::new(Int32Array::from(vec![1, 2])),
+                Arc::new(StringArray::from(vec!["a", "b"])),
+            ],
+        )
+        .unwrap();
+
+        let reader = Box::new(arrow_array::RecordBatchIterator::new(
+            vec![Ok(batch1)],
+            schema.clone(),
+        ));
+
+        let source_table = db
+            .create_table(CreateTableRequest {
+                name: "independent_source".to_string(),
+                namespace: vec![],
+                data: CreateTableData::Data(reader),
+                mode: CreateTableMode::Create,
+                write_options: Default::default(),
+            })
+            .await
+            .unwrap();
+
+        let source_uri = db.table_uri("independent_source").unwrap();
+
+        // Clone the table
+        let cloned_table = db
+            .clone_table(CloneTableRequest {
+                target_table_name: "independent_clone".to_string(),
+                target_namespace: vec![],
+                source_uri,
+                source_version: None,
+                source_tag: None,
+                is_shallow: true,
+            })
+            .await
+            .unwrap();
+
+        // Both should start with 2 rows
+        assert_eq!(source_table.count_rows(None).await.unwrap(), 2);
+        assert_eq!(cloned_table.count_rows(None).await.unwrap(), 2);
+
+        // Add data to the cloned table
+        let batch_clone = RecordBatch::try_new(
+            schema.clone(),
+            vec![
+                Arc::new(Int32Array::from(vec![3, 4, 5])),
+                Arc::new(StringArray::from(vec!["c", "d", "e"])),
+            ],
+        )
+        .unwrap();
+
+        let cloned_table_obj = Table::new(cloned_table.clone());
+        cloned_table_obj
+            .add(Box::new(arrow_array::RecordBatchIterator::new(
+                vec![Ok(batch_clone)],
+                schema.clone(),
+            )))
+            .execute()
+            .await
+            .unwrap();
+
+        // Add different data to the source table
+        let batch_source = RecordBatch::try_new(
+            schema.clone(),
+            vec![
+                Arc::new(Int32Array::from(vec![10, 11])),
+                Arc::new(StringArray::from(vec!["x", "y"])),
+            ],
+        )
+        .unwrap();
+
+        let source_table_obj = Table::new(source_table.clone());
+        source_table_obj
+            .add(Box::new(arrow_array::RecordBatchIterator::new(
+                vec![Ok(batch_source)],
+                schema.clone(),
+            )))
+            .execute()
+            .await
+            .unwrap();
+
+        // Verify they have evolved independently
+        assert_eq!(source_table.count_rows(None).await.unwrap(), 4); // 2 + 2
+        assert_eq!(cloned_table.count_rows(None).await.unwrap(), 5); // 2 + 3
+    }
+
+    #[tokio::test]
+    async fn test_clone_latest_version() {
+        let (_tempdir, db) = setup_database().await;
+
+        // Create a source table with initial data
+        let schema = Arc::new(Schema::new(vec![Field::new("id", DataType::Int32, false)]));
+
+        let batch1 =
+            RecordBatch::try_new(schema.clone(), vec![Arc::new(Int32Array::from(vec![1, 2]))])
+                .unwrap();
+
+        let reader = Box::new(arrow_array::RecordBatchIterator::new(
+            vec![Ok(batch1)],
+            schema.clone(),
+        ));
+
+        let source_table = db
+            .create_table(CreateTableRequest {
+                name: "latest_version_source".to_string(),
+                namespace: vec![],
+                data: CreateTableData::Data(reader),
+                mode: CreateTableMode::Create,
+                write_options: Default::default(),
+            })
+            .await
+            .unwrap();
+
+        // Add more data to create new versions
+        for i in 0..3 {
+            let batch = RecordBatch::try_new(
+                schema.clone(),
+                vec![Arc::new(Int32Array::from(vec![i * 10, i * 10 + 1]))],
+            )
+            .unwrap();
+
+            let source_table_obj = Table::new(source_table.clone());
+            source_table_obj
+                .add(Box::new(arrow_array::RecordBatchIterator::new(
+                    vec![Ok(batch)],
+                    schema.clone(),
+                )))
+                .execute()
+                .await
+                .unwrap();
+        }
+
+        // Source should have 8 rows total (2 + 2 + 2 + 2)
+        let source_count = source_table.count_rows(None).await.unwrap();
+        assert_eq!(source_count, 8);
+
+        let source_uri = db.table_uri("latest_version_source").unwrap();
+
+        // Clone without specifying version or tag (should get latest)
+        let cloned_table = db
+            .clone_table(CloneTableRequest {
+                target_table_name: "cloned_latest".to_string(),
+                target_namespace: vec![],
+                source_uri,
+                source_version: None,
+                source_tag: None,
+                is_shallow: true,
+            })
+            .await
+            .unwrap();
+
+        // Cloned table should have all 8 rows from the latest version
+        assert_eq!(cloned_table.count_rows(None).await.unwrap(), 8);
+    }
+}

rust/lancedb/src/index/vector.rs

@@ -8,7 +8,7 @@
 //! values
 use std::cmp::max;
 
-use lance::table::format::{Index, Manifest};
+use lance::table::format::{IndexMetadata, Manifest};
 
 use crate::DistanceType;
 
@@ -19,7 +19,7 @@ pub struct VectorIndex {
 }
 
 impl VectorIndex {
-    pub fn new_from_format(manifest: &Manifest, index: &Index) -> Self {
+    pub fn new_from_format(manifest: &Manifest, index: &IndexMetadata) -> Self {
         let fields = index
             .fields
             .iter()

rust/lancedb/src/remote/db.rs

@@ -14,9 +14,9 @@ use serde::Deserialize;
 use tokio::task::spawn_blocking;
 
 use crate::database::{
-    CreateNamespaceRequest, CreateTableData, CreateTableMode, CreateTableRequest, Database,
+    CloneTableRequest, CreateNamespaceRequest, CreateTableData, CreateTableMode,
-    DatabaseOptions, DropNamespaceRequest, ListNamespacesRequest, OpenTableRequest,
+    CreateTableRequest, Database, DatabaseOptions, DropNamespaceRequest, ListNamespacesRequest,
-    TableNamesRequest,
+    OpenTableRequest, TableNamesRequest,
 };
 use crate::error::Result;
 use crate::table::BaseTable;
@@ -27,6 +27,18 @@ use super::table::RemoteTable;
 use super::util::{batches_to_ipc_bytes, parse_server_version};
 use super::ARROW_STREAM_CONTENT_TYPE;
 
+// Request structure for the remote clone table API
+#[derive(serde::Serialize)]
+struct RemoteCloneTableRequest {
+    source_location: String,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    source_version: Option<u64>,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    source_tag: Option<String>,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    is_shallow: Option<bool>,
+}
+
 // the versions of the server that we support
 // for any new feature that we need to change the SDK behavior, we should bump the server version,
 // and add a feature flag as method of `ServerVersion` here.
@@ -430,6 +442,51 @@ impl<S: HttpSend> Database for RemoteDatabase<S> {
         Ok(table)
     }
 
+    async fn clone_table(&self, request: CloneTableRequest) -> Result<Arc<dyn BaseTable>> {
+        let table_identifier = build_table_identifier(
+            &request.target_table_name,
+            &request.target_namespace,
+            &self.client.id_delimiter,
+        );
+
+        let remote_request = RemoteCloneTableRequest {
+            source_location: request.source_uri,
+            source_version: request.source_version,
+            source_tag: request.source_tag,
+            is_shallow: Some(request.is_shallow),
+        };
+
+        let req = self
+            .client
+            .post(&format!("/v1/table/{}/clone", table_identifier.clone()))
+            .json(&remote_request);
+
+        let (request_id, rsp) = self.client.send(req).await?;
+
+        let status = rsp.status();
+        if status != StatusCode::OK {
+            let body = rsp.text().await.err_to_http(request_id.clone())?;
+            return Err(crate::Error::Http {
+                source: format!("Failed to clone table: {}", body).into(),
+                request_id,
+                status_code: Some(status),
+            });
+        }
+
+        let version = parse_server_version(&request_id, &rsp)?;
+        let cache_key = build_cache_key(&request.target_table_name, &request.target_namespace);
+        let table = Arc::new(RemoteTable::new(
+            self.client.clone(),
+            request.target_table_name.clone(),
+            request.target_namespace.clone(),
+            table_identifier,
+            version,
+        ));
+        self.table_cache.insert(cache_key, table.clone()).await;
+
+        Ok(table)
+    }
+
     async fn open_table(&self, request: OpenTableRequest) -> Result<Arc<dyn BaseTable>> {
         let identifier =
             build_table_identifier(&request.name, &request.namespace, &self.client.id_delimiter);
@@ -1221,4 +1278,146 @@ mod tests {
             _ => panic!("Expected Runtime error from header provider"),
         }
     }
+
+    #[tokio::test]
+    async fn test_clone_table() {
+        let conn = Connection::new_with_handler(|request| {
+            assert_eq!(request.method(), &reqwest::Method::POST);
+            assert_eq!(request.url().path(), "/v1/table/cloned_table/clone");
+            assert_eq!(
+                request.headers().get("Content-Type").unwrap(),
+                JSON_CONTENT_TYPE
+            );
+
+            let body = request.body().unwrap().as_bytes().unwrap();
+            let body: serde_json::Value = serde_json::from_slice(body).unwrap();
+            assert_eq!(body["source_location"], "s3://bucket/source_table");
+            assert_eq!(body["is_shallow"], true);
+
+            http::Response::builder().status(200).body("").unwrap()
+        });
+
+        let table = conn
+            .clone_table("cloned_table", "s3://bucket/source_table")
+            .execute()
+            .await
+            .unwrap();
+        assert_eq!(table.name(), "cloned_table");
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_with_version() {
+        let conn = Connection::new_with_handler(|request| {
+            assert_eq!(request.method(), &reqwest::Method::POST);
+            assert_eq!(request.url().path(), "/v1/table/cloned_table/clone");
+
+            let body = request.body().unwrap().as_bytes().unwrap();
+            let body: serde_json::Value = serde_json::from_slice(body).unwrap();
+            assert_eq!(body["source_location"], "s3://bucket/source_table");
+            assert_eq!(body["source_version"], 42);
+            assert_eq!(body["is_shallow"], true);
+
+            http::Response::builder().status(200).body("").unwrap()
+        });
+
+        let table = conn
+            .clone_table("cloned_table", "s3://bucket/source_table")
+            .source_version(42)
+            .execute()
+            .await
+            .unwrap();
+        assert_eq!(table.name(), "cloned_table");
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_with_tag() {
+        let conn = Connection::new_with_handler(|request| {
+            assert_eq!(request.method(), &reqwest::Method::POST);
+            assert_eq!(request.url().path(), "/v1/table/cloned_table/clone");
+
+            let body = request.body().unwrap().as_bytes().unwrap();
+            let body: serde_json::Value = serde_json::from_slice(body).unwrap();
+            assert_eq!(body["source_location"], "s3://bucket/source_table");
+            assert_eq!(body["source_tag"], "v1.0");
+            assert_eq!(body["is_shallow"], true);
+
+            http::Response::builder().status(200).body("").unwrap()
+        });
+
+        let table = conn
+            .clone_table("cloned_table", "s3://bucket/source_table")
+            .source_tag("v1.0")
+            .execute()
+            .await
+            .unwrap();
+        assert_eq!(table.name(), "cloned_table");
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_deep_clone() {
+        let conn = Connection::new_with_handler(|request| {
+            assert_eq!(request.method(), &reqwest::Method::POST);
+            assert_eq!(request.url().path(), "/v1/table/cloned_table/clone");
+
+            let body = request.body().unwrap().as_bytes().unwrap();
+            let body: serde_json::Value = serde_json::from_slice(body).unwrap();
+            assert_eq!(body["source_location"], "s3://bucket/source_table");
+            assert_eq!(body["is_shallow"], false);
+
+            http::Response::builder().status(200).body("").unwrap()
+        });
+
+        let table = conn
+            .clone_table("cloned_table", "s3://bucket/source_table")
+            .is_shallow(false)
+            .execute()
+            .await
+            .unwrap();
+        assert_eq!(table.name(), "cloned_table");
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_with_namespace() {
+        let conn = Connection::new_with_handler(|request| {
+            assert_eq!(request.method(), &reqwest::Method::POST);
+            assert_eq!(request.url().path(), "/v1/table/ns1$ns2$cloned_table/clone");
+
+            let body = request.body().unwrap().as_bytes().unwrap();
+            let body: serde_json::Value = serde_json::from_slice(body).unwrap();
+            assert_eq!(body["source_location"], "s3://bucket/source_table");
+            assert_eq!(body["is_shallow"], true);
+
+            http::Response::builder().status(200).body("").unwrap()
+        });
+
+        let table = conn
+            .clone_table("cloned_table", "s3://bucket/source_table")
+            .target_namespace(vec!["ns1".to_string(), "ns2".to_string()])
+            .execute()
+            .await
+            .unwrap();
+        assert_eq!(table.name(), "cloned_table");
+    }
+
+    #[tokio::test]
+    async fn test_clone_table_error() {
+        let conn = Connection::new_with_handler(|_| {
+            http::Response::builder()
+                .status(500)
+                .body("Internal server error")
+                .unwrap()
+        });
+
+        let result = conn
+            .clone_table("cloned_table", "s3://bucket/source_table")
+            .execute()
+            .await;
+
+        assert!(result.is_err());
+        if let Err(crate::Error::Http { source, .. }) = result {
+            assert!(source.to_string().contains("Failed to clone table"));
+        } else {
+            panic!("Expected HTTP error");
+        }
+    }
 }

rust/lancedb/src/table.rs

@@ -1976,6 +1976,8 @@ impl NativeTable {
     /// Delete keys from the config
     pub async fn delete_config_keys(&self, delete_keys: &[&str]) -> Result<()> {
         let mut dataset = self.dataset.get_mut().await?;
+        // TODO: update this when we implement metadata APIs
+        #[allow(deprecated)]
         dataset.delete_config_keys(delete_keys).await?;
         Ok(())
     }
@@ -1986,6 +1988,8 @@ impl NativeTable {
         upsert_values: impl IntoIterator<Item = (String, String)>,
     ) -> Result<()> {
         let mut dataset = self.dataset.get_mut().await?;
+        // TODO: update this when we implement metadata APIs
+        #[allow(deprecated)]
         dataset.replace_schema_metadata(upsert_values).await?;
         Ok(())
     }