public interface DataRecordUpload
|Modifier and Type||Method and Description|
The largest part size the client can send in a multi-part upload.
The smallest part size the client can send in a multi-part upload (not counting the final part).
Returns a token that uniquely identifies this upload.
Returns a Collection of direct-writable upload URIs for uploading a file, or file part in the case of multi-part uploading.
@NotNull @NotNull String getUploadToken()
Note that some backends have lower-bound limits for the size of a part of a multi-part upload.
A smaller size may also be used so long as it exceeds the value returned
getMinPartSize(). Such smaller values may be more desirable
for clients who wish to tune uploads to match network conditions;
however, the only guarantee offered by the API is that using parts of
the size returned by
getMaxPartSize() will work without using
more URIs than those available in the Collection returned by
If a client calls
DataRecordAccessProvider.initiateDataRecordUpload(long, int) with a
maxUploadSizeInBytes that ends up being smaller than the
actual size of the binary to be uploaded, it may not be possible to
complete the upload with the URIs provided. The client should initiate
the transaction again with the correct size.
Note that some backends have upper-bound limits for the size of a part of a multi-part upload.
@NotNull @NotNull Collection<URI> getUploadURIs()
DataRecordAccessProvider.initiateDataRecordUpload(long, int), OR - If the implementing data store does not support multi-part uploading, OR - If the client-specified value for maxUploadSizeInBytes in a call to
DataRecordAccessProvider.initiateDataRecordUpload(long, int)is less than or equal to the minimum size of a multi-part upload part If the collection contains only a single URI the client should treat that URI as a direct single-put upload and write the entire binary to the single URI. Otherwise the client may choose to consume up to the entire collection of URIs provided.
Note that ordering matters; URIs should be consumed in sequence and not skipped.
Copyright © 2012–2020 The Apache Software Foundation. All rights reserved.