KNIME 4.1.0 (64-bit)

What's new in this version:

New node:
- Send to Power BI
- Google BigQuery Connector
- Google Cloud Storage (2 nodes)
- Create Databricks Environment
- Databricks File System Connection
- Integration of Amazon Personalize services
- MDF Reader -- Measurement Data Format
- Split Density Scorer into Updater and Scorer
- Add Active Learning Loop
- H2O Integration: Add H2O MOJO Predictor (Isolation Forest) node
- Excel Writer (new file handling)
- Excel Reader (new file handling)
- Line Reader (new file handling)
- Weak Label Model Learner and Predictor
- Run R Model in Microsoft SQL Server (new db framework)
- Nominal Probability Distribution Creator and Splitter
- DB Connection Closer
- DB Connection Extractor
- Amazon Athena Connector (new)
- Webpage Retriever
- Silhouette Coefficient
- Top K Selector Node
- Row Filter (Labs) -- supporting multiple, more advanced search criteria
- Vertica Connector (new)

Enhancement:
- Ensure a menu item for the Node Monitor in the View menu
- Add option to Boolean Widget and Configuration nodes to push flow variable of type boolean
- Select background when opening a component
- Add success message after successfull deployment of preferences
- Component metadata editor
- Show DB Meta Data Browser in DB Table Selector when custom query is enabled
- Make input ports of concatenate node configurable
- Add visual indicator for nodes whose ports are configurable
- Show "Recent Workflows" under "File" menu
- Pass NodeCreationConfiguration to createNodeDialogPane
- Allow to disable DB Data port in DB Manipulation nodes
- Support none local KNIME URL resolving in JDBC parameters
- Allow earlier garbage collection of cached tables
- Allow garbage collection of in-memory tables that are being iterated over
- Text Processing: Arabic language support
- Enable flow variables of type boolean, long, array for Table Row/Column to Variable and Variable to Table Row/Column node
- Implement new File Handling Port Type
- Implement Nominal Probability Distributions with Metadata
- Support Probability Distribution Values in Uncertainty Scorer Nodes
- Modify NodeFactoryClassMapper to expose mappings
- Suggest URL template for known db types
- Component editor with example data
- Probability Distribution Creator: One-hot encoding of String columns
- Create Temp Directory node should create directory inside workflow
- Implement responsive design for welcome page
- Support TLS1.2 in the Send Email node
- Label model: Calculate covariance matrix in KNIME
- Support probabilistic Inputs in the Label Model
- Table Reader Node to have timeout option (important when reading from remote resources)
- Improved JDBC URL template UI
- DB transaction management including Start and End Transaction node
- Rank/Linear Correlation: save p-value and degrees of freedom in correlation model
- Configurations and Widgets: improve error message on variable name starting with "knime"
- Support loading of a native library per user defined JDBC driver
- Cache static model contents of density scorer models
- Untangle Testing Framework feature
- Let Label Model output probabilistic labels
- Workflow migration tool tooltip
- Write code-base allowing to add/remove ports to/from standard nodes
- Allow multiple table views within one view
- CSV Writer: Make comment header indentation optional
- Switching to "Welcome Page" should clear Description View
- Restrict metadata editor link types via pulldown
- Rewrite Potential Uncertainty Scorer Node
- Custom Renderer for Probabilistic Labels
- Probability Distribution Cell
- Revise Exploration/Exploitation Scorer (Rename to Exploration/Exploitation Score Combiner)
- Rewrite Graph Density Uncertainty Scorer Node
- Enhance Logistic Regression to support a probabilistic target column
- Enhance Gradient Boosted Trees to support a probabilistic target column
- Revise Uncertainty Scorer Nodes
- Change default settings of the Partitioning nodes
- Add flow variable support to DB Table Selection component
- Drag'n'drop of workflow from Hub
- Allow to disable DB connection recreation on workflow loading
- Improve node description of the Feature Selection Loop Start nodes
- Improved error messaging on Components (e.g. when contained nodes are not fully connected or fail on execution)
- Remove annotation border menu and toolbar items for node annotations
- H2O Integration: Support MOJOs of category 'Ordinal'
- H2O Integration: Add option to treat unknown categorical values as NA to MOJO predictors
- Rank/Linear Correlation: show p-values and number of cases for each correlation
- Improve usability of Zooming
- Add more space to the workbench to the bottom and right (additional space when moving or adding nodes)
- JavaScript Line Plot with option to start x and y axis at 0 and not at minimum
- H2O Integration: Improve tool tips of H2O ports and labels/layout in H2O frame view
- K-Means: Add option to randomly select initial centroids
- List Remote Files: Update Node Description
- PMML Predictor should output probabilities and allow change of prediction column name
- Additional flow variable types for boolean, long, and arrays
- (Big Data Extensions): Allow custom configuration keys in custom Spark settings of the Livy node
- (Big Data Extensions): Support user name and password authentication in Create Spark Context (Livy) node e.g. to connect to Azure HDInsight
- (Big Data Extensions): Support Livy 0.6 in KNIME
- (Big Data Extensions): Persistently save and display Kerberos username
- (Big Data Extensions): Implement fallback solution for ORC/Parquet reader
- (Big Data Extensions): Add support for KNOX to Create Spark Context (Livy) node
- (Big Data Extensions): Add WebHDFS via KNOX Connector node
- (Big Data Extensions): Add support for KNOX to Hive Connector node

Bug Fixes:
- Unable to parse container.compression property Gzip (on system with Turkish Locale)
- PCA node fails when large number of input columns are present
- SSH Connection does not resolve ~ on linux updating to 4.0
- KNIME doesn't open with empty workspace (CentOS 6)
- Naive Bayes Model handles derived fields incorrectly
- PMML Transformation Applier fails when no TransformationDictionary is present
- PMML Transformation Applier fails to apply missing value and/or normalization model
- Slider Widget loads with JS error
- Metanode disappears when converting to component
- Right click on empty canvas does not deselect component
- NullPointerException when closing Parquet Table Store Reader
- Math Formula (Multi Column) does not seem to work with certain column names
- Git Push node: Push tags checkbox in config dialog not configurable
- Column/Variable Expressions: Function selection dialog stays open after closing configuration dialog
- Column Filter configuration validation does not import options created in 4.0
- NullPointerException when loading Workflow with Excel Reader node that reads from Amazon S3
- Writing files to custom URL fails
- Close component metadata editor when left
- Edit "icon" missing when stepping into component after it has been unlinked
- Excel Writer has option to save file relative to workflow, but does not seem to support it
- Column Filter Configuration filters out all columns if the dialog is never opened
- Loading org.knime.ext.ssh bundle slows KNIME start drastically (class traversal issue)
- Gradient Boosted Trees to PMML produces invalid PMML in some cases
- Support setting alpha and beta to 0 for Tversky Distance
- DB manipulation nodes (Insert, Update, Merge, Delete, Writer) do not work in loops
- BarChart highlights wrong row on selection
- Range Slider Filter Definition Node has rounding errors
- Flow Variable should be spelled correctly in the node port type
- Node annotation toolbar doesnt disappear when switching windows
- Custom URL does not work with file scheme
- DB Reader throws duplicate key exception when reading multiple binary columns
- Typo in Update Components dialog
- Pie/Donut Chart can't be configured if Number type column is not present in table
- CSS styling not working for SVG output of Plotly nodes
- Random Forest progress seems stuck during the creation of the final trees
- Choosing KNIME FileSystem causes the Dialog to resize so that Browse Button is not visible
- Windows Eclipse 4.7 does not support horizontal scrolling via scrollwheel
- Missing Value node description incorrect
- Rank Correlation: Rank output should be of type long not double
- Scroll bar resizes during scrolling in H2O and Spark table views for large datasets
- When selecting for the axis column in the Line Plot (Plotly) it fails when executing
- Wrong port color displayed for BufferedDataTable ports on components
- Table Difference Checker ignores domain order
- Mouse goes to upper left corner during loading of streamable nodes
- Color legend too small on Plotly Bubble Chart node
- New database feature added to knime-ap target platform
- Drag'n'drop problem with Image Processing nodes from Hub to AP
- Update File Reader node description
- Python: stderr messages should be logged to INFO unless they're warnings
- DB Connector node: Database URL (if long) is not moving together with the cursor to the end of the URL
- KNIME types without a mapping rule are shown in the Source Type dropdown box
- Stacked Area Chart assigns colors based on column order instead of column name
- Executed Keras Input Layer doesn't work after upload and download from server
- Support cancellation of fetch unique values query in DB Row Filter
- Remove duplicates in node repository
- Streamable Metanode can be reset during execution
- Problem restoring file store cells on re-loaded of executed and streamed component
- Empty Table Switch fails in streamed node
- Streaming Wrapped nodes fails with NPE in inactive Branch
- (Big Data Extensions): DB Loader for Hive has problems with input KNIME tables that have non-word characters in their name
- (Big Data Extensions): H2O nodes do not show string columns on Sparkling Water