Changelog for KNIME Analytics Platform 3.1

Changes since 3.1.2

(only available on update site --- not part of product builds)

  • [AP-3828] - Potential ConcurrentModificationException while configuring downstream nodes after execution
  • [AP-5521] - Statistic Node fails with IndexOutOfBounds Exception
  • [AP-5689] - KNIME hangs in splash screen if internet connection is broken
  • [AP-5839] - R nodes can't populate flow variables
  • [AP-5840] - Extract JavaScript resources for view creation also from plugins which are not extracted (directly from JAR)
  • [AP-5848] - No updates found although some are available / "Expected timestamp" error message when installing new extensions
  • [AP-5907] - Wrapped Metanodes to automatically hide "knime.*" flow variable constants -- problems when used on KNIME server
  • [AP-5915] - Difference Checker node cannot handle collections with missing values
  • [AP-5931] - Unlocking of "Wrapped Metanode" not properly working
  • [AP-5937] - Creating a target platform via New Node Wizard fails if path contains spaces
  • [AP-5971] - Twitter Users Node: Matching of rows to results is case sensitive while Twitter results are not
  • [AP-5982] - JSON Reader uses system charset instead of UTF-8
  • [AP-6005] - Increase startup timeout to work around slow virus scanners
  • [AP-6025] - JSON Input node always returns default input values as input parameters in REST response


Changes between 3.1.1 and 3.1.2

Release date: April 1, 2016

  • [AP-5643] - Windows installer not recognized as trusted application any more
  • [AP-5644] - Wait for queued AWT events before closing workflows during tests
  • [AP-5646] - R views may fail on Mac when XQuartz is missing
  • [AP-5649] - NodeMonitor not unregistered from node when view is closed
  • [AP-5654] - Missing Value fails for Most Frequent Value when input is Integer
  • [AP-5657] - The PMML Model Appender has problems with transformations when dealing with ensembles
  • [AP-5661] - The PMML Ensemble Predictor cannot deal with models that rely on global transformations
  • [AP-5667] - Filestores created in wrapped metanodes, which are part of loops, can't be saved
  • [AP-5680] - Rule Engine doesn't accept "fake" String Cells
  • [AP-5685] - Table To JSON should not translate booleans to numbers
  • [AP-5688] - Streamable node filter in node repository has problems
  • [AP-5698] - DialogComponentStringListSelection should not pre-select items if no selection is required
  • [AP-5704] - SSH nodes are missing in some installations
  • [AP-5708] - LIKE in Rule Engine node doesn't handle multiline strings
  • [AP-5713] - Node Repository: Uncaught error while searching for streamable nodes
  • [AP-5716] - NullPointerException in Copy/Move node when location doesn't have a protocol
  • [AP-5717] - Parso (SAS reader) library cannot handle certain SAS7BDAT files
  • [AP-5728] - DateAndTimeValue without milliseconds produces error when writing to database table
  • [AP-5729] - Applying XPaths may drop namespace declarations
  • [AP-5730] - Missing Value (Apply) node does not fill in median in an integer column if the median is of type double
  • [AP-5735] - Table To PMML Ensemble does not merge MiningSchemas correctly
  • [AP-5744] - File Meta Info node has incorrect node description
  • [AP-5745] - MySQL driver no longer able to write array types into database
  • [AP-5746] - Text Mining: Stop Word Filter should support knime:// protocol
  • [AP-5748] - R snippets to better handle headless execution (disable pop-outs when no display connection exist)
  • [AP-5756] - Body nodes of parallel chunk loops cannot be streamed (only 'top level' wrapped metanodes)
  • [AP-5757] - Database Query node removes \ and $ from input query
  • [AP-5760] - Database GroupBy should insert extra 'AS' statement for Google BigQuery (not fully SQL compliant)
  • [AP-5761] - The PMML Model Appender fails for non-Decision Tree models
  • [AP-5784] - Credentials validation in all database connect nodes happens too early during load phase
  • [AP-5811] - Equality checker outputs wrong expected and actual values in case of failure

Changes between 3.1.0 and 3.1.1

Release date: February 1, 2016

  • Enh 6603: Update free Marvin Chemistry extension to 3.1.0.v0501
  • Enh 6634: New extension point for custom actions before the first workflow is loaded (feature request from customer)
  • Enh 6721: Option to evaluate XLS-formulas on writing (feature request from customer)
  • Enh 6658: Text Mining: Word Parser should read docm files
  • Enh 6744: Test framework: Add optional include pattern to Unittestrunner
Bug Fixes:
  • Bug 6680: NullPointer on saving autobinner model in "Model to Cell" node
  • Bug 6684: Try-Catch Nodes don't populate error message in "fail" branch
  • Bug 6127: Column list loop start fails if input column names change (should ignore)
  • Bug 6621: Used Python command is only refreshed after user hit Apply or OK in Preferences, not if it is changed programmatically
  • Bug 6625: KMeans fails when input column is integer
  • Bug 6631: R Integration: ggplot2 graphs are not drawn when subsequent commands follow
  • Bug 6623: Python: 'Restore Defaults' button in preference page does not use preference defaults
  • Bug 6712: SDF Insert node dialog only allows selecting "RowID" as molecule title (should be all eligible columns)
  • Bug 6713: Metanodes cause error during execution of workflow in wrapped node wizard execution mode
  • Bug 6637: Statistics Node - potential (misleading) AssertionError when loading executed node
  • Bug 6619: Better error messages when Rserve is not installed
  • Bug 6644: Potential deadlock in WorkflowManager during report generation in KNIME server executor
  • Bug 6708: Text Mining: Stop Word Filter ignores custom stop word file
  • Bug 6735: Browsing example workflows from welcome page doesn't work any more
  • Bug 6734: New Quickforms should only be visible in wrapped metanode dialog
  • Bug 6715: k-means: cluster table (new in 3.1.0) has inconsistent cluster naming
  • Bug 6656: NPE in explorer browser after incorrect configuration
  • Bug 6738: XL Reader: No error message if file has unsupported format
  • Bug 6740: Documentation of "String to Date" node misses some new field options (e.g. 'Y' - Week Year)
  • Bug 6742: R Integration: Rserve processes should be launched with ProcessBuilder (inherits environment better)
  • Bug 6743: Wrong default R home may be detected
  • Bug 6655: NPE in SQLLiteConnector when file is on server.
  • Bug 6650: Weka Dialogs cannot open Choose for Classifier
  • Bug 6739: Weka Predictor is too restrictive when checking the test and model column types
  • Bug 6737: XLS Reader shows exception in log when a new file is set
  • Bug 6757: Problems with empty workflowset.meta files
  • Bug 6756: KNIME occasionally hangs during start under some Linux systems


Changes between v3.0.1 and v3.1.0
(see highlight summary for KNIME Analytics Platform v3.0 and v3.1)

Release date: December 6, 2015

New Nodes:

  • New Node 6344: Database Numeric Binner
  • New Node 6596: Database Auto-Binner
  • New Node 6060: Database Sampling
  • New Node 5652: Database Pivoting
  • New Node 6582: Active Learning (Labs) - multiple new nodes
  • New Node 6521: Rank: Derive Rank for values in a column
  • New Node 5789: Reference Row & Column Splitter
  • New Node 3673: Credentials input quickform node
  • New Node 6052: Twitter node to extract retweets
  • New Node 5940: Twitter Followers Search node
  • New Node 1791: Simple Regression Tree Learner + Predictor
  • New Node 6182: Random Forest Learner/Predictor for both classification and regression
  • New Node 6474: Prediction Fusion (combine predictions from different classifiers)
  • New Node 6486: Improved Enrichment Plotter


  • Enh 6562: Added adapter cell implementation for more molecule types
  • Enh 6577: Restructure update sites
  • Enh 6497: Add Java Cryptography Extension to bundled JRE
  • Enh 6402: "Modular PMML" to be part of standard KNIME distribution (not labs) - all PMML learner & preproc node now without optional PMML input
  • Enh 6421: Update Apache CXF web service library to 3.0.7
  • Enh 6495: Update RSyntaxTextArea to version 2.5.8
  • Enh 6351: R Integration: Independent nodes to run in parallel (using RServe as opposed to RJava) - used to be statically sync'ed for data exchange
  • Enh 6565: Active Branch Inverter with optional input (is currently mandatory)
  • Enh 6420: Logistic Regression supporting bit/byte vector types
  • Enh 6405: Many preprocessing and predictor nodes to follow streaming API (~70 additional nodes)
  • Enh 6429: New context menu entry for WrappedNodes: "Unwrap"
  • Enh 6419: Text Processing nodes to support streamable API
  • Enh 6445: JSON editor for Subnode layouts
  • Enh 6432: Parallel Chunker loop construct lets status of workflow alternate between "executing" and "marked" (potential inconsistency when run on server)
  • Enh 6120: Scorer node with option to better deal with missing values
  • Enh 6444: Define JSON structure for Bootstrap based Subnode layout
  • Enh 6568: Ability to filter node repository for streamable nodes
  • Enh 6569: Better search functionality in node repository (fuzzy search)
  • Enh 6132: Tree Ensemble Node to support byte vector data type
  • Enh 5493: Text Mining: PDF Parser and Word Parser should allow file names as document titles
  • Enh 4847: K-Means cluster node to export cluster centers as separate (2nd) output

Bug Fixes:

  • Bug 6572: Text Mining: NPE in Tagger nodes for empty input tables
  • Bug 6571: Text Mining: Update openNLP to 1.6
  • Bug 6485: ColumnAppender node with additional options to make it less restrictive (don't just wrap the table), e.g. allow different rowIDs
  • Bug 6559: Renderer to Image node allows selecting renderers in dialog that are not available
  • Bug 5223: Text Mining: General Tag Filter does not filter out terms with certain tags correctly.
  • Bug 6583: Streaming Executor (Labs) does not handle finished "consumer" nodes correctly
  • Bug 6507: GroupBy node should be less strict when checking column types
  • Bug 6604: PMML Compiler has problem with Logistic Regression models
  • Bug 5486: KNIME Welcome page opens in external browser window on some systems
  • Bug 6520: Parso (SAS reader) library cannot handle certain SAS7BDAT files
  • Bug 6523: Molecule string renderers are not aware of adapter cells
  • Bug 6524: Text Mining: BoW Creator need missing value check on document col
  • Bug 6525: Text Mining: Reduce default setting for number of threads in Strings to Document node
  • Bug 6536: Streaming Executor (Labs) doesn't properly release file store objects (memory leak)
  • Bug 5338: Text Mining: Document grabber returns empty table if document directory does not exist
  • Bug 6527: Text Mining: Strings to Document streamable
  • Bug 6537: (API) New framework method: ExecutionContext#createConcatenateTable - efficient concatenation of tables
  • Bug 6530: Text Mining: ABNER tagger is not thread safe
  • Bug 6540: Concatenate Node: Use ConcatenateTable instead of AppendedRowsTable
  • Bug 6533: Use Cmd+double click on Mac for editing wrapped nodes
  • Bug 6534: Resetting multiple table view nodes at once may lead to deadlock
  • Bug 4995: Text Mining: Meta Info Extractor changes Row Ids
  • Bug 5773: Text Mining: RowIDs should start with 0 not with 1
  • Bug 6516: UI Artifacts since 3.0 (follow up to bug 6414)
  • Bug 6423: Potential NPE in ContainerTable (only seen in test cases)
  • Bug 6053: Location and Tweet ID support for the Twitter nodes
  • Bug 6551: Weka Predictor and Cluster Assigner unnecessarily copy data and keep data in memory
  • Bug 5499: R interactive nodes don't kill R process when cancelled.
  • Bug 5687: KNIME Crash while running simple transpose in R (with 50k columns)
  • Bug 5937: Text Mining: TaggerCellFactory need missing value check
  • Bug 6532: org.knime.chem.base need to export sdfwriter2 package
  • Bug 6340: Full-Full KNIME Build with all extensions, including server-side extension and client-connector
  • Bug 6408: Review node description XSD to allow more expressive elements for description texts
  • Bug 6416: Streaming Executor (Labs) doesn't save data when workflow is written/restored
  • Bug 6401: Review default font size choosing (hardcoded to 8 in product --- which is often too small)
  • Bug 6561: The Wait node does not accept the knime protocoll and need progress status
  • Bug 6417: WorkflowManager may sent too many "unstable" change events - should defer to when workflow lock is released
  • Bug 6567: Wrong testcase name when test workflow is zipped
  • Bug 6438: Weka 3.7 Predictor crashes if Regression is performed on Integer Target Column
  • Bug 6456: Weka nodes: Multiple occurring options weren't saved properly
  • Bug 6400: Test Framework: Assertion Log may let test cases fail (although uncriticial) - backport from 3.0
  • Bug 6461: Dialog of Java Snippet is missing rearrangement options
  • Bug 6073: DialogComponentColumnFilter2 throws exception when changing search pattern in dialog
  • Bug 6457: Try-Catch nodes behaves strange when error occurs in a side-branch (correct error message is missing)
  • Bug 6483: Weka-Node Costsensitive Classification has problems loading settings
  • Bug 6490: List Box Input does not allow dot as separator
  • Bug 6489: Weka: Rework the way the options of a weka-classifier, clusterer etc. are stored in the dialog settings
  • Bug 6491: 'Workspace is locked' message contains '{0}' - should be workspace path

What are you looking for?