Load data in multiple parts by using the Cloud Object Storage API

With the Cloud Object Storage API, you can load data objects as large as 5 GB in a single PUT, and objects as large as 5 TB by loading the data into object storage as a set of parts which can be loaded independently in any order and in parallel. After all of the parts have been loaded, they are presented as a single object in Cloud Object Storage.

You can load files with these formats and mime types in multiple parts:

• application/xml
• application/pdf
• text/plain; charset=utf-8

To load a data object in multiple parts:

1. Initiate a multipart load:

curl -X "POST" "https://(endpoint)/(bucket-name)/(object-name)?uploads"
 -H "Authorization: bearer (token)"

The values for bucket-name and token are on the project's General page on the Manage tab. Click Manage in IBM Cloud on the Watson Studio for the endpoint value.

2. Load the parts by specifying arbitrary sequential part numbers and an UploadId for the object:

curl -X "PUT" "https://(endpoint)/(bucket-name)/(object-name)?partNumber=(sequential-integer)&uploadId=(upload-id)"
 -H "Authorization: bearer (token)"
 -H "Content-Type: (content-type)"

Replacecontent-type with application/xml, application/pdf or text/plain; charset=utf-8.

3. Complete the multipart load:

  curl -X "POST" "https://(endpoint)/(bucket-name)/(object-name)?uploadId=(upload-id)"
 -H "Authorization: bearer (token)"
 -H "Content-Type: text/plain; charset=utf-8"
 -d \$'<CompleteMultipartUpload>
         <Part>
           <PartNumber>1</PartNumber>
           <ETag>(etag)</ETag>
         </Part>
         <Part>
           <PartNumber>2</PartNumber>
           <ETag>(etag)</ETag>
         </Part>

4. Add your file to the project as an asset. From the Assets page of your project, click the Upload asset to project icon. Then, from the Files pane, click the action menu and select Add as data set.

Next steps

• Refining the data
• Analyzing the data and building models

Learn more

• Storing very large objects in Cloud Object Storage
• Using curl to store very large objects

Parent topic: Adding connections to projects