KNIME 4.4.0 (64-bit) 軟體 Download 下載

• 1

  版本確認

• 2

  檔案準備中

• 3

  檔案下載

What's new in this version:

Fixed:
- Table to PDF & HTML cannot write pngs
- Twitter Streaming node failure
- Move file name column to the end rather than prepending it in reader nodes
- NPE in Parameter Optimization Loop Start
- Annotation text on Linux sometimes not readable (font not installed)
- Add better error message for SPA not working in Bundled Chromium and SWT Browser
- Workflow Writer / Deploy to Server nodes do not work if the server URL contains a port and contains a trailing slash
- Unusual / blurry node rendering on macOS
- Community converter can overwrite KNIME converter
- MacOS Big Sur: Various screen artifacts (distorted table, broken splash screen, etc...)
- Conda Environment Propagation Dialog forever loads if Python is not set up
- Fix Locales for extract date & time
- Python: File extension mapping of pickled files (.pkl) maps to deprecated Python Object Reader node
- "Workflow Executor" exports variables in reverse order
- Race condition (and memory leak) when data is retrieved and evicted from the cache at the same time
- Conda Environment Propagation: Kernel queue is not cleared upon failure during environment creation
- Python: Deprecated port type selectable in component setup
- Column selection set to "Document" instead of "Preprocessed Document" in workflow example
- Add data to Reference Reader nodes recursively (for workflow ports)
- Domain Calculator throws exception if Path column is included
- NPE on DnD of an URI lacking a host
- Problem with connections reaching into a capture scope whose source is the parent workflow (i.e. metanode port)
- Fix virtual scope context retrieval from metanode
- Parallel Chunk Loop should retain port objects which lead into a capture scope for later usage
- TRF based Readers don't recalculate the spec if spec affecting settings are controlled via flow variable
- Wrapping/scrolling of status message doesn't work properly
- FileReader throws wrong exception text
- SettingsModelReader/Writer show filter options though they are not available
- Workflow Writer displays confusing warning on escaped default workflow name
- TensorFlow 2 Reader: Reading a network in SavedModel format from TensorFlow Hub fails with ProviderMismatchException
- Check if “Custom workflow name” in ”Capture Workflow End“ node has illegal characters
- Problem with credentials flow variables when injected via reference reader node
- Workflow temp directory deleted while workflow is running
- Configure in Joiner slow for wide tables
- Send to Power BI node: Upload with special column name fails