軟體下載 KNIME 4.2.0 (64-bit)

• 1

  版本確認

• 2

  檔案準備中

• 3

  檔案下載

What's new in this version:

New:
- Joiner (Labs) -- multiple outputs, performance, UX, ...
- Integrated Deployment Nodes -- Capture Workflow Start/End, Workflow Combiner, Workflow Writer, ...
- Various new reader and writer nodes supporting bulk reading and remote file system connection (Labs)
- Microsoft Authentication
- SAP Reader (based on Theobald Xtract Universal)
- SharePoint Online Connector
- Simple File Reader -- simplified usage, faster execution time
- TensorFlow 2 Keras Reader
- Google Cloud Storage Connector (Labs)
- String Manipulation (Multiple Columns)
- Create Temp Dir node
- Row Splitter (Labs)
- List Files/Folders (Labs) (new file handling)
- Nodes for Amazon DynamoDB
- Table Difference Finder
- Salesforce Connection (multiple nodes, e.g. Salesforce (OAuth2) Connector, SOQL Query, ...)

Enhancement:
- Widget nodes: New design in component views (old mode available via 'legacy' flag in layout editor)
- Add workflow data area to the "Relative To" category (new file handling)
- "Spotfire (Labs)" nodes deprecated -- fully replaced by partner contributed nodes
- Textprocessing Performance Improvements
- Enable file text field in new file handling dialog in remote workflow editor
- Update Java to JRE to version "8u252-b09"
- Update Jackson to 2.11.0
- LineReader - Add BOM and gz support
- File menu entry to export a workflow summary
- Add duplicate checker to RowKey column
- Preference Page for Off-Heap Column Store activation
- urlDecode String Manipulator
- Register UrlEncode String Manipulator with Column Expressions Node
- Next functionality for node search
- String Manipulation node(s): new 'urlencode' method
- Support component drag'n'drop from private spaces (hub) into AP
- Colorize the connection lines of a selected node
- Metanode fails with "Error in sub flow"
- Support exchangeability of port types
- Create FSLocation Variable Type
- Table View (output port) with new option to copy data with header (to paste into Excel)
- Make 'Number to String' node streamable
- Disable dyn. port adaptations during execution and adopt warning message
- Make R environment configurable on a node level
- Add dynamic port support for Merge Variables node
- Update KNIME Analytics Platform to use the latest Eclipse (Eclipse 2020-03, i.e. 4.15)
- Add option to export summary / provenance trace from executed workflow
- Display column headers options should be added to the labeling view
- Make Column Appender with flexible Number of Input ports
- Add 'Configuration', 'Widget' and 'Container' node types
- Provide Ungroup node to skip rows with empty collection
- Top K Selector to give all occurrences of n biggest/smallest values
- Add portobject to possible porttypes of components
- Update Sorter UI for Sorter and top-k Node
- Add Node Monitor as default view
- Add float64 (double) datatype in Keras Input Layer
- Change default search mode in Table view to be case insensitive and to include column names
- Workflow Test Framework to Test for Errors in Components
- Add option to replace included columns in Column Combiner node
- Add select best option to the Feature Selection Filter node
- Allow to open dialog of DB nodes with missing/invalid input connection
- Support reconnecting to database if connection is invalid
- Improve formatting of long flow variable names in flow variable tab
- Implement SWT.OpenUrl event to support opening KNIME URLs for all OSs
- DB Loader: Support Microsoft SQL Server bulk copy API
- Support boolean columns in DB Row Filter
- Inform user about absence of flow variables in nodes with no settings
- Hide memory policy tab in nodes with no data table output ports
- Missing Value: Support fixed boolean value
- CSV Reader: Reduce number of opened input streams
- New 'hidden'-flag for node(set) extensions
- (Big Data Extensions): Add note about cluster-level permissions to node description of Create Databricks Environment
- (Big Data Extensions): Support Timezone option in Create Spark Context nodes

Bug Fixes:
- Perl integration does not work with Perl >=5.30 any more
- Provide fallback for non-standard SQL types when reading from databases
- Flow Variables don't work on Salesforce Node
- H2O Integration: Local context issues on certain IPsec VPN setups
- Amazon S3 Connection (Labs) sometimes does not list all subdirectories of a directory
- Java Snippet does not clear warning messages on reset
- Tableau Hyper API: Node fail when data volume increases.
- Text Mining: Add invertible flag to StanfordNLP Tokenizers to ensure tokens not to be normalized
- Support non-standard JDBC Oracle types such as BINARY_FLOAT and BINARY_DOUBLE
- Parallel Chunk Start dialogue has bugs: missing flow variable and always greyed out checkbox
- Scorer node output column named specifity instead of specificity
- PortTypeRegistry fails ungracefully when encountering erroneous port types among all available port types
- REST nodes don't work with empty passwords in BASIC authentication
- Amazon Athena Connector node could not login without role switching
- Temporary table and file stores in loops not deleted on clear
- Table Creator: Copying of cells with cmd+c doesn't work on Mac
- Null pointer exception in Table Column to Variable node
- Wrong workflow svg stored when saving workflow while in annotation edit mode
- OutputStreams can remain open after workflow was closed
- Empty folders in workflow directory are missing from export
- Workflow editor locks up when saving large workflow using temporary workflow editor
- GET Request node: inconsistency with retrieved binary objects
- Math Formula: Check for column names containing dollar signs
- Fix in-workflow buffers not being cleared when workflow is closed and cleaned up
- Deadlock when using flow variable buttons with JTextFields
- Inner nested components duplicate error messages
- Word metanode still used when checking if update available
- Configuration of Call Workflow (Table Based) node fails, if wf on server contains File Download node in wrapped metanode
- Make dropping node on connections and nodes a uniform experience
- Executing follow up nodes block reset of non-related content of metanode
- Unused Metanode input causes warnings upon workflow opening
- FTP node's open() always calls changeDirectoryUp at least one time
- Make "Lift Chart (JavaScript)" node responsive
- Execute option to work on all nodes
- (Big Data Extensions): Parquet Reader and ORC Reader fail when reading directories with hidden files
- (Big Data Extensions): Create Local Big Data Environment node fails to create root scratch directory on Windows if domain server is not accessible