Overview
Uploading local files to a volume allows you to upload local files of any format (images, audio/video, PDFs, model weights, raw logs, and so on) to a Volume under Catalog as raw bytes, preserving the original file structure. After the upload is complete, the files are visible in the Data Catalog as Volume resources and can be consumed by downstream consumers such as Notebooks, Python jobs, and Agents. Dimension | Upload local files to a table | Upload local files to a volume |
File Processing Method | Parse files into structured fields by file type | Store as raw bytes |
File Type | Only certain file types | Any format |
Whether Computing Resources Are Required | Yes (for parsing) | No |
Maximum Number of Files per Upload | 10 | 100 |
Maximum Size per File | (Total size limit: 2 GB) | 5 GB (per file) |
Target Location | Catalog → Schema → Table | Catalog → Schema → Volume |
Applicable Scenarios
Machine learning sample ingestion: Upload labeled images, audio, and text corpora as original files to a Volume for direct reading and training by Notebooks.
Models and configuration files: Upload model artifacts such as model weights (.pt / .pkl) and configuration files (.yaml).
Stage raw logs: Place the raw logs to be parsed into a Volume first, and then parse them into a table through a Python job in a Workflow.
Image and document archiving: Upload unstructured assets such as PDFs, contracts, and scanned documents in batches.
Operation Steps
Step 1: Entering File Upload to Volume
In the console, go to the Data Integration module. On the Access Method Tab, click the File Upload to Volume card. The system displays the File Upload to Volume dialog box.
Step 2: Selecting a File
Two methods are supported:
Drag and drop: Drag files directly to the upload area.
Click to select: Click the Select File button to select one or more files in the system file browser.
Constraint item | Limit | Validation message |
Number of files | At least 1, up to 100 | You can upload up to 100 files. |
Single File Capacity | ≤ 5 GB | A single file cannot exceed 5 GB. |
File Type | Any format | Extension name validation is not performed. |
Complete error list:
Scenario | Note. |
File quantity exceeded | Upload failed. You can upload up to 100 files. |
Single file exceeds size limit. | Upload failed. A single file cannot exceed 5 GB. |
Network disconnected. | Upload failed. The network connection is disconnected. |
Permission changed. | Upload failed. No permission to perform the upload service. |
Fallback | Upload failed. {Error details}. |
After files are added to the list, you can:
Click X to remove a single file.
Click Select File again to add more files.
Step 3: Selecting a Destination Path
Parameter | Description | Required |
Target Path | Select the target Volume. Only Volume type directories can be selected. If no Volume exists in the current Catalog / Schema, create one immediately in the tree directory. | Yes |
To create a Volume:
Hover over the Schema. A + icon appears on the right. Click it to open the creation panel.
Tip
The Volume creation action is independent of the current upload process. Even if you cancel this upload after creating a Volume, the created Volume remains in the Catalog.
Parameter | Description | Required |
Same-Name Handling | Auto rename (default): generate a new file name file name_1; replace existing file: overwrite the existing file with the same name with the latest data. | Yes |
Only Catalogs / Schemas / Volumes on which the current account has write permission are displayed. You can use the search box to fuzzy match directory names.
Step 4: Uploading
Click Upload:
The system returns to the Data Integration homepage and displays the Upload Progress List in the lower-right corner.
The progress percentage of each file is displayed as a background bar in the list.
Operations supported by the progress list:
Operation | Description |
Go to resource. | After the upload is complete, click the link icon next to the file to go to the Volume details page where the file resides. You cannot go to the page before the upload is complete. |
Hide / expand list. | Click the close icon in the upper-right corner to hide the list as a collapsed progress indicator at the top of the page. Click the indicator to expand the list again. |
Step 5: Verifying the Result
After the upload is complete, you can:
1. Navigate to the target Volume in Data Catalog and confirm that the file exists. 2. Read the Volume file by its file path in Notebook. 3. Use a Python File task in Workflow to read and process files in the Volume. FAQs
Q: What should I do if a single upload exceeds 100 files or 5 GB?
A: Upload data in multiple batches. If a large amount of raw data is stored in COS or a file system, use File to Volume Sync to ingest data through periodic offline sync tasks. Q: Can files uploaded to a volume be parsed back into structured tables?
A: Yes. After uploading, you can create a SQL task or Python task to read Volume files and write them to structured tables. You can also use File to Table Sync to load files from the Volume into tables in batches. Q: What happens if I close the browser during upload?
A: Incomplete uploads will be interrupted. When you enter the page again, files that have been successfully uploaded to the Volume are not affected, while interrupted files need to be uploaded again.
Q: Can an upload be resumed after a network interruption?
A: Resumable upload is not supported. If a "Network connection disconnected" message appears, wait for the network to recover and then upload the file again.
Q: Why can't I select the destination path?
A: The target Volume path must at least reach the volume node (the full path is catalog > schema > volume > file directory). You can click the small triangle on the left side of the file tree to expand it until the volume node or file directory node is reached.
References