GridFSFileWriter

GridFSFileWriter provides a streaming interface for uploading large files to GridFS. It handles chunking the file data and managing the upload process, including error handling and cleanup.

Basic Usage

// Create a writer
let writer = try await GridFSFileWriter(
    toBucket: gridFS,
    fileId: ObjectId(), // Optional custom ID
    chunkSize: 261_120  // Optional custom chunk size (default: 255KB)
)

// Write data in chunks
for chunk in dataChunks {
    try await writer.write(data: chunk)
}

// Finalize and create the file
let file = try await writer.finalize(
    filename: "large-file.dat",
    metadata: [
        "contentType": "application/octet-stream",
        "description": "Important data"
    ]
)

Streaming from HTTP

let writer = try await GridFSFileWriter(toBucket: gridFS)

do {
    // Stream file from HTTP request
    for try await chunk in request.body {
        try await writer.write(data: chunk)
    }

    // Complete the upload
    let file = try await writer.finalize(
        filename: "uploaded-file.dat"
    )
} catch {
    // Clean up partial upload
    try await writer.cancel()
    throw error
}

Error Handling

• If an error occurs during upload, call cancel() to clean up partial chunks

• The writer becomes invalid after calling finalize() or cancel()

• Writing to a finalized writer will trigger an assertion failure

Performance Tips

• The default chunk size (255KB) is suitable for most use cases

• Larger chunks reduce the number of database operations but use more memory

• The writer buffers data until it has a full chunk before writing to GridFS

• Call flush() to force writing a partial chunk to the database

Implementation Details

• Each chunk is stored as a separate document in the chunks collection

• Chunks are numbered sequentially starting from 0

• The file metadata is only written when finalize() is called

• Indexes are automatically created on the first write to a bucket