This is an automated email from the ASF dual-hosted git repository.
kou pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow.git
The following commit(s) were added to refs/heads/main by this push:
new 068416bd41 GH-46296: [Swift] Add support for reading struct (#46302)
068416bd41 is described below
commit 068416bd411d6a8e2949f8ebcb2f80e2c302ef6b
Author: abandy <[email protected]>
AuthorDate: Sat May 3 17:35:24 2025 -0400
GH-46296: [Swift] Add support for reading struct (#46302)
### Rationale for this change
`makeArrayHolder` supports `ArrowTypeId.strct`.
https://github.com/apache/arrow/blob/62c59c23352583980e894a43ec46335c4d55a7e1/swift/Arrow/Sources/Arrow/ArrowReaderHelper.swift#L189-L190
However, we cannot use `makeArrayHolder` if `fromProto` fails to handle
`org_apache_arrow_flatbuf_Type_.struct_`.
https://github.com/apache/arrow/blob/62c59c23352583980e894a43ec46335c4d55a7e1/swift/Arrow/Sources/Arrow/ArrowReaderHelper.swift#L139-L148
### What changes are included in this PR?
This PR aims to fix `ProtoUtil.fromProto` to support `struct_`.
This PR also adds an unit test for it.
This is based on GH-46297.
### Known issues with Swift Arrow Reader
Currently the Swift Arrow reader requires a null buffer to be included in
the list of buffers. The reader will fail if the null buffer is missing.
### Are these changes tested?
Yes.
### Are there any user-facing changes?
No.
* GitHub Issue: #46296
Co-authored-by: Dongjoon Hyun <dongjoon@ apache.org>
Authored-by: Alva Bandy <[email protected]>
Signed-off-by: Sutou Kouhei <[email protected]>
---
swift/Arrow/Sources/Arrow/ArrowType.swift | 2 ++
swift/Arrow/Sources/Arrow/ProtoUtil.swift | 2 ++
swift/Arrow/Tests/ArrowTests/IPCTests.swift | 40 ++++++++++++++++++++++++++++-
swift/data-generator/swift-datagen/go.mod | 3 ++-
swift/data-generator/swift-datagen/main.go | 28 ++++++++++++++++++++
5 files changed, 73 insertions(+), 2 deletions(-)
diff --git a/swift/Arrow/Sources/Arrow/ArrowType.swift
b/swift/Arrow/Sources/Arrow/ArrowType.swift
index 000fe21568..b44f859185 100644
--- a/swift/Arrow/Sources/Arrow/ArrowType.swift
+++ b/swift/Arrow/Sources/Arrow/ArrowType.swift
@@ -385,6 +385,8 @@ extension ArrowType.Info: Equatable {
return lhsId == rhsId
case (.timeInfo(let lhsId), .timeInfo(let rhsId)):
return lhsId == rhsId
+ case (.complexInfo(let lhsId), .complexInfo(let rhsId)):
+ return lhsId == rhsId
default:
return false
}
diff --git a/swift/Arrow/Sources/Arrow/ProtoUtil.swift
b/swift/Arrow/Sources/Arrow/ProtoUtil.swift
index f7fd725fe1..ac61030c08 100644
--- a/swift/Arrow/Sources/Arrow/ProtoUtil.swift
+++ b/swift/Arrow/Sources/Arrow/ProtoUtil.swift
@@ -64,6 +64,8 @@ func fromProto( // swiftlint:disable:this
cyclomatic_complexity
let arrowUnit: ArrowTime64Unit = timeType.unit == .microsecond ?
.microseconds : .nanoseconds
arrowType = ArrowTypeTime64(arrowUnit)
}
+ case .struct_:
+ arrowType = ArrowType(ArrowType.ArrowStruct)
default:
arrowType = ArrowType(ArrowType.ArrowUnknown)
}
diff --git a/swift/Arrow/Tests/ArrowTests/IPCTests.swift
b/swift/Arrow/Tests/ArrowTests/IPCTests.swift
index 311a3bb235..4f56f5fdab 100644
--- a/swift/Arrow/Tests/ArrowTests/IPCTests.swift
+++ b/swift/Arrow/Tests/ArrowTests/IPCTests.swift
@@ -118,7 +118,44 @@ func makeRecordBatch() throws -> RecordBatch {
}
}
-final class IPCFileReaderTests: XCTestCase {
+final class IPCFileReaderTests: XCTestCase { // swiftlint:disable:this
type_body_length
+ func testFileReader_struct() throws {
+ let fileURL =
currentDirectory().appendingPathComponent("../../testdata_struct.arrow")
+ let arrowReader = ArrowReader()
+ let result = arrowReader.fromFile(fileURL)
+ let recordBatches: [RecordBatch]
+ switch result {
+ case .success(let result):
+ recordBatches = result.batches
+ case .failure(let error):
+ throw error
+ }
+
+ XCTAssertEqual(recordBatches.count, 1)
+ for recordBatch in recordBatches {
+ XCTAssertEqual(recordBatch.length, 3)
+ XCTAssertEqual(recordBatch.columns.count, 1)
+ XCTAssertEqual(recordBatch.schema.fields.count, 1)
+ XCTAssertEqual(recordBatch.schema.fields[0].type.info,
ArrowType.ArrowStruct)
+ let column = recordBatch.columns[0]
+ XCTAssertNotNil(column.array as? StructArray)
+ if let structArray = column.array as? StructArray {
+ XCTAssertEqual(structArray.arrowFields?.count, 2)
+ XCTAssertEqual(structArray.arrowFields?[0].type.info,
ArrowType.ArrowString)
+ XCTAssertEqual(structArray.arrowFields?[1].type.info,
ArrowType.ArrowBool)
+ for index in 0..<structArray.length {
+ if index == 2 {
+ XCTAssertNil(structArray[index])
+ } else {
+ XCTAssertEqual(structArray[index]?[0] as? String,
"\(index)")
+ XCTAssertEqual(structArray[index]?[1] as? Bool, index
% 2 == 1)
+ }
+ }
+ }
+
+ }
+ }
+
func testFileReader_double() throws {
let fileURL =
currentDirectory().appendingPathComponent("../../testdata_double.arrow")
let arrowReader = ArrowReader()
@@ -384,3 +421,4 @@ final class IPCFileReaderTests: XCTestCase {
}
}
}
+// swiftlint:disable:this file_length
diff --git a/swift/data-generator/swift-datagen/go.mod
b/swift/data-generator/swift-datagen/go.mod
index c645e6759b..ca0fd5446e 100644
--- a/swift/data-generator/swift-datagen/go.mod
+++ b/swift/data-generator/swift-datagen/go.mod
@@ -16,7 +16,8 @@
module swift-datagen/main
-go 1.22.7
+go 1.23.0
+
toolchain go1.24.1
require github.com/apache/arrow-go/v18 v18.2.0
diff --git a/swift/data-generator/swift-datagen/main.go
b/swift/data-generator/swift-datagen/main.go
index 97b807d9bb..dbe24222fa 100644
--- a/swift/data-generator/swift-datagen/main.go
+++ b/swift/data-generator/swift-datagen/main.go
@@ -82,7 +82,35 @@ func writeDoubleData() {
writeBytes(rec, "testdata_double.arrow")
}
+func writeStructData() {
+ mem := memory.NewGoAllocator()
+
+ fields := []arrow.Field{
+ {Name: "my struct", Type: arrow.StructOf([]arrow.Field{
+ {Name: "my string", Type: arrow.BinaryTypes.String},
+ {Name: "my bool", Type: arrow.FixedWidthTypes.Boolean},
+ }...)},
+ }
+
+ schema := arrow.NewSchema(fields, nil)
+
+ bld := array.NewRecordBuilder(mem, schema)
+ defer bld.Release()
+
+ sb := bld.Field(0).(*array.StructBuilder)
+ f1b := sb.FieldBuilder(0).(*array.StringBuilder)
+ f2b := sb.FieldBuilder(1).(*array.BooleanBuilder)
+
+ sb.AppendValues([]bool{true, true, false})
+ f1b.AppendValues([]string{"0", "1", ""}, []bool{true, true, false})
+ f2b.AppendValues([]bool{false, true, false}, []bool{true, true, false})
+
+ rec := bld.NewRecord()
+ writeBytes(rec, "testdata_struct.arrow")
+}
+
func main() {
writeBoolData()
writeDoubleData()
+ writeStructData()
}
