RoboForm

What's new in this version:

New nodes:
- Delete SharePoint Online List
- SharePoint Online List Writer
- DB Looping
- DB Delete (Filter)
- DB Data Spec Extractor
- DB Concatenate
- UIEXT-327: Echarts "Scatter Plot"
- UIEXT-190: Table View
- UIEXT-6: ECharts “Bar chart”
- UIEXT-5: ECharts "Line plot"

Enhancements:
- NXT-1144: KNIME Modern UI Preview - preview of the new user interface
- Linux: Add warning popup in case Wayland display server is detected (standard in Ubuntu 22.04. and currently not properly supported)
- Provide Python API to develop KNIME nodes
- JSON To XML node with option to remove invalid characters
- Regex Splitter Node: Doesn't handle multiline strings
- Call Workflow Service to use the Mountpoint Connector
- Update Apache CXF from 3.2.14 to 3.5.2
- Help menu with additional entry redirecting to KNIME Forum
- "Transpose" node to be memory-aware, auto-determining batch size
- Allow to extend Python path programmatically in nodes that use the Python3KernelBackend
- Columnar Backend: Reorder virtual operations to evaluate "slice" before "append"
- Case Switch End Node: Execution time significantly increased when only one input is active (pass on input table)
- Add support for using a custom Azure AD application with the Microsoft Authentication node
- Structure XGBoost Learner settings according to tabs
- Update of log4j-1.2.15 to 1.2.21 (reload4j)
- Support Path type in Variable Expressions node
- H2 driver update to version 2.1.212
- Amazon S3 Connector should use multi-part upload for large files
- Support path flow variables in Column Expressions node
- Add option to specify base score in XGBoost Learners
- Improve Python Script (Labs) node's Python process connection startup time
- Add row key to error messages in XGBoost nodes
- Update H2O to 3.36.0.3
- Store tag sets as part of DataColumnSpec
- Add colsample_bynode option in XGBoost Tree Ensemble Learner
- Allow to set global thread pool size via environment variable
- Microsoft SQL Server node to support transaction disabling to be compatible with Serverless SQL Pool in Azure Synapse Analytics
- Allow setting+appending values of extension types in Pandas Dataframe
- Polling configuration (advanced settings) for Call Workflow Service node
- MySQL driver updated to version 8.0.29
- PostgreSQL driver updated to version 42.3.5
- Amazon Redshift driver updated to version 2.1.0.3
- Provide Oracle Database driver version 19.14.0 as new feature for installation
- Allow keeping RowIDs when joining
- Show hub host name in tab
- SQLite driver updated to version 3.36.0.3
- Support "NOT LIKE" in DB Row Filter
- Allow to select from all available dialects for default db type
- Microsoft Access driver updated to version 5.0.1
- Static bar values in Bar Chart
- Use bundled Python environment in Python Script (Labs) node
- Let Missing Cell directly override isMissing() of DataCell
- Column Expression Node: Ability to access previous and next rows
- XGBoost: Update to version 1.5.1
- Support row weights in XGBoost
- Add linear model parameters as additional output for the XGBoost Linear Ensemble Learner
- Feature Importance Output for XGBoost Tree Ensemble Learner
- XGBoost: Support for vector (bit/byte/double) features
-Generic Web Service Client node (SOAP) to use recent versions of CXF client library (3.5.2)
- (Big Data Extensions): Update H2O Sparkling Water to 3.36.1.2
- (Big Data Extensions): Add support for H2O Sparkling Water to 3.32.1.2-1-3.1
- (Big Data Extensions): Add Spark 3.1 support to Create Databricks Environment node
- (Big Data Extensions): Add Spark 3.1 support to Create Spark Context (Livy) node
- (Big Data Extensions): Add support for Spark 3.1 to KNIME Extension for Apache Spark
- (Big Data Extensions): Update H2O Sparkling Water to 3.36.0.3-1-3.2
- (Big Data Extensions): Add Spark 3.2 support to Create Databricks Environment node
- (Big Data Extensions): Update Create Local Big Data Environment to Spark 3.2.1
- (Big Data Extensions): Add support for Spark 3.2 to KNIME Extension for Apache Spark
- (Big Data Extensions): Add support for new type annotations to Parquet Writer (new file handling)
- NXT-1072: Remote workflow editor: enable configuration nodes in component dialog

Fixed:
- R integration does not work with R 4.2.0 any more (on Ubuntu)
- Fixed Width File Reader: Column settings are deleted when controlling the location via flow variable on opening the dialog
- Excel Writer generates error if overwriting a sheet that is used as source in pivot table
- Variable Condition Loop End error is missing information (only log output)
- Table Reader has problems with BlobDataCells when reading multiple tables
- REST Nodes don't work with encoded query parameter names
- "Call Workflow" nodes do not hide any system credentials when invoking remote service (causing error)
- BufferedRowWrite#setMissing is off by one
- The MDF Reader feature doesn't properly install Python
- Read Images Node: Fails on SVG input
- Component is not ignoring error message caught by Try Catch
- Installation directory on Windows must not be writable by everyone
- Copying table cells out of table view node is not working in CEF
- Image Port Type: Unable to clone input data invalid character
- Webpage Retriever fails at cookie and error cause extraction
- Check for correct conda package versions by Python (Labs) in the Pref Page
- Conda Environment Propagation blocks Executor on Configure
- PythonGateway creation gets stuck because std out and error are not consumed
- Inconsistent Behaviour between Match All and Match Any when comparing Row IDs in Joiner
- Parquet/ORC Writer: Mode File/Folder configuration not displayed correctly
- Breakpoint node description is not showing available settings
- File Download Widget: Download Link not working if Component ID is 0
- Cache size heuristics do not make sense on MacOS
- Fix missing value handling in Lists of Pandas ExtensionTypes
- JDBC type REAL looses precision due to double conversion
- Keyboard shortcut for updating linked component does not work on top-level workflow
- Save as workflow doesn't show all spaces on MacOs
- Make sure we do not install jedi=0.18.1 in conda envs because it is broken
- Hard JVM crashes on Windows caused by XGBoost nodes run in parallel
- Writing to the root folder in SharepointOnline causes a NullPointerException
- HTTP(S) Connector always fails to trust bad SSL certificates
- KNIME UI sometimes not coming up in Linux Dev environment
- reg:linear is deprecated in new XGBoost version
- Case Switch End node changes its initial configuration if dialog is opened
- DB Writer fails in streaming mode with enabled DB Data outport
- Integrated Deployment: Variables in "static" input are injected in reverse order
- DB Transaction End/Catch node with inactive input ports does not return inactive spec in component
- KNIME AP freezes when several metanodes have the same input connection
- Workflow Executor does not work when workflow has no inputs/outputs
- XGBoost: Linear Booster Updater 'Shotgun' cannot handle some selectable feature selectors
- Nullpointer after closing KNIME AP due to unfinished wf coach loading
- "Reset" workflow action resets nodes with executing successors
- Documentation of JSON Processing extension talks about XML
- Improve error message when empty tables are sent to XGBoost
- Workflow Summary Extractor extracts and stores passwords
- Python Script node prohibits deletion of jobs folder on executor
- Webpage Retriever: NPE on certain webpages
- Inconsistent capitalization of items in Help menu
- Reader/Writer nodes with Encoding tab should not be initialized to use the OS default encoding option
- DataCells are not cached in the heap cache
- Update smbj library in SMB Connector to potentially fix errors (NullPointerException, PipeShare error) when running parallel operations against DFS
- Mean calculated with Column Aggregator node is inconsistent with other methods
- Node Dialog too wide with long Flow Variable list
- Send Email node does not encode Chinese characters in attachment file name
- Feature buffer empty exceptions in XGBoost Learner and Predictor
- Sunburst chart has error after selection, close and reopen
- XGBoost can lead to memory problems and KNIME crashes
- XGBoost Predictor fails on unknown categorical values
- Itemset Miner (Borgelt) uses too much memory by default
- Rename XGBoost Linear Ensemble Learner