[ARROW-10719] [C#] ArrowStreamWriter doesn't write schema metadata

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 3.0.0
• Component/s: C#
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/26667
• Assignee: Unassigned
• Reporter: Eric Erhardt
• Created: 24/Nov/20 19:42
• Updated: 11/Jan/23 08:15
• Resolved: 16/Dec/20 03:42

Description

See this TODO in the code:

https://github.com/apache/arrow/blob/master/csharp/src/Apache.Arrow/Ipc/ArrowStreamWriter.cs#L445

 

A simple test case:

 

// Build schema
var schemaBuilder = new Schema.Builder();
schemaBuilder.Field(f => f
    .Name("Ints")
    .Nullable(true)
    .DataType(Int32Type.Default)
    .Metadata(new Dictionary<string, string> { { "packed", packed ? "true" : "false" } })
);
var indexMetadata = new Dictionary<string, string> { { "index", JsonConvert.SerializeObject(new object()) } };
var schema = schemaBuilder.Metadata(indexMetadata).Build();
Int32Array a = new Int32Array.Builder().Append(5).Append(4).Build();
// Create record batch
var recordBatch = new RecordBatch(schema, new IArrowArray[] { a }, a.Length);// Write recordbatch to file
using (var stream = File.Create("visits.arrow"))
using (var writer = new ArrowFileWriter(stream, recordBatch.Schema))
{
    writer.WriteRecordBatch(recordBatch);
    writer.WriteEnd();
}using (var stream = File.OpenRead("visits.arrow"))
using (var reader = new ArrowFileReader(stream))
{
    var batch = reader.ReadNextRecordBatch();
    Console.WriteLine(batch.Schema.HasMetadata);
}

The above should print `true`, but it prints `false` for HasMetadata.

Issue Links

• links to: GitHub Pull Request #8902