KNIME Analytics Platform 3.7.2
Release date: April 18, 2019
Enhancements:
- AP-11412: DevDoc: Write a new KNIME Node Extension Quickstart Guide
- AP-11261: (API) Add hook to workflow loading routine to populate "knime.system.credentials"
- AP-9412: Add new regions to AWS S3 Connector node
Bug Fixes:
- AP-11596: Import preferences does not properly import mountpoints
- AP-11531: Changing Multi-line editor width in String input Quickform does not work in 3.7
- AP-11488: H2O confuses column name order when used with Spark Context
- AP-11399: Error when input table is empty for "Row To Variable Loop Start" node
- AP-11370: H2O Integration: Dialog of Table to H2O node does not work anymore
- AP-11367: Column Enforce Inclusion/Exclusion Bug in Multiple Learner Nodes
- AP-11583: POST Request node 'request body' configuration problems
- AP-11580: H2O Integration: Typo in "H2O Multinominal Scorer"
- AP-11577: KNIME Table API: RowIteratorBuilder not 'future proof' (will be removed from API and redone in 3.8)
- AP-11530: Tile View (JavaScript) does not display SVGs in WebPortal
- AP-11525: Stacked Area Chart assigns colors based on column order instead of column name
- AP-11479: Network Viewer (JavaScript) Nullpointer Exception
- AP-11466: Post Request nodes forgets URL settings on load
- AP-11459: Send to Tableau Server (Hyper) node is not performing overwrites correctly
- AP-11364: Dead link in OPTICS Cluster Assigner node description
- AP-11359: 'POST Request node' - settings can throw java exceptions
- AP-11356: Early stopping of Random Search does not work properly in certain cases
- AP-11347: H2O scorer nodes fail with unmeaningful error messages if reference column contains missing values
- AP-11342: Continued SWT leak hunting
- AP-11317: JavaScript Histogram View fails on missing values
- AP-11271: Missing Whiskers in Boxplot (JavaScript)
- AP-11155: Java Snippet node configuration dialog loads awfully slow
- BD-905: (Big Data Extensions): Hive to Spark/Spark to Hive do not work on HDP 3.x
- BD-901: (Big Data Extensions): Adding flow variables in PySpark node breaks preset text areas
- BD-900: (Big Data Extensions): Spark to H2O fails on 'Short' columns
- BD-844: (Big Data Extensions): Fix column spec domain in Table to H2O Node with Sparkling Water Context
- BD-838: (Big Data Extensions): Spark to H2O node only works with H2O Sparkling Water Context, configure fails with bad error message
KNIME Analytics Platform 3.7.1
Release date: January 29, 2019
Enhancements:
- AP-11007: DL Framework: Add simple summary view to DL Port Objects
- AP-10984: Rename "Card View" node to "Tile View" (otherwise infringing trademark held by 3rd party)
- AP-10560: Update deprecated node of Cross Validation metanode
Bug Fixes:
- AP-11264: Sorting table in table view causing documents to be shown as missing values
- AP-11212: FileChooser doesn't work in WebPortal with Firefox
- AP-11181: Refreshing parts of the workspace after workflow rename take longer than in 3.6
- AP-11174: R integration: Columns are not kept if an empty table is returned from R to KNIME
- AP-11032: Mountpoint gets deleted upon restart if it was set to inactive
- AP-11031: Expression Engine: Scrollbar does not resize/appear
- AP-10959: Expression Engine: min, max doesn't support Date&Time anymore
- AP-11222: H2O Integration: Scorer node mixes up columns in the confusion matrix
- AP-11192: Map Oracle type Date to local date time and not local date
- AP-11182: Row Filter framework might cause problems during UI component initialization
- AP-11162: "Table Row to Variable Loop Start": Table iterator leak making KNIME slower and slower for large number of iterations/tables
- AP-11151: More SWT Windows handle leakage
- AP-11124: DL Keras: Recurrent layers are broken
- AP-11063: Cell Splitter by Position node checks wrong settings for column name duplicates during configuration
- AP-11055: XGBoost Learner: Column moves from excluded to included on dialog reopen
- AP-11049: Workflow editor missing after trying to maximize workflow
- AP-11042: H2O Predictor nodes do not set a default column name in the dialog as other Predictor nodes do
- AP-11041: DL Network Executor fails in streaming mode with Round Double node
- AP-11039: Clean up dangling SWT handles
- AP-11019: RegexMatcher in Expressions returns String
- AP-11018: Text Mining: Tag Filter does not work properly with Zemberek Tags
- AP-10982: Open workflow fails if db session for executed nodes can not be restored and thus does not exist
- AP-10975: Tensorflow Extensions claims ownership of 'org.knime.base' -- nodes are possibly associated with this feature in the node description/usage records
- AP-10947: R version detection reports invalid installation when Rserve (or any other dependency) is missing
- AP-10912: DB Connector node should update type mapping when database type changes
- AP-10891: Send Email waiting endless for email server response
- AP-10735: DL TensorFlow Reader: Error console output during typing
- BD-883: (Big Data Extensions): Spark Entropy Scorer fails with NullPointerException on missing values
- BD-878: (Big Data Extensions): PySpark Snippet nodes fail in local big data environment on Windows
- BD-776: (Big Data Extensions): Unpersist Spark DataFrame/RDD node makes DataFrame/RDD unavailable
- BD-655: (Big Data Extensions): Spark GroupBy fails on numeric aggregations for boolean columns
- BD-654: (Big Data Extensions): Spark Scorer should ignore rows with missing values
- BD-653: (Big Data Extensions): Spark Collaborative Filtering Learner dialog should should not allow boolean columns as user or product columns
- BD-630: (Big Data Extensions): Spark Category to Number fails on missing values
KNIME Analytics Platform 3.7.0
(see highlight summary)
Release date: December 6, 2018
New Nodes:
- AP-10977: XGBoost Integration: Various new nodes integrating the XGBoost machine learning library (Labs)
- AP-10435: Google Authentication (interactive authentication for Google Drive/Sheets/Analytics/...)
- AP-8967: Updated Tableau Integration ('Hyper' support -- two nodes: Write Hyper file and Send to Server)
- AP-7812: Histogram (JavaScript)
- AP-10123: Card View (JavaScript)
- AP-7811: Hierarchical Cluster Assigner (JavaScript)
- AP-5181: Heatmap (JavaScript)
- AP-10036: Kolmogorov-Smirnov Test (statistical hypothesis test)
- AP-9010: CSS Editor
- AP-10976: Database Integration (Labs): Rewrite of existing DB client nodes, supporting more connectors (e.g. Oracle, SQL Server), connection pooling, rich type support, powerful editors, etc. (part of KNIME Labs)
- AP-9938: Friedman Test (non-parametric statistical test)
- AP-9937: Proportion Test (statistical hypothesis test)
- AP-9907: Container Input (Credentials)
- AP-9466: Google Drive Connection
- AP-10404: Turkish POS Tagger
- AP-10403: Turkish Stemmer
- BD-781: (Big Data Extensions): PySpark Script nodes
- BD-435: (Big Data Extensions): Spark Row Filter node
Enhancements:
- AP-10809: Upgrade bundled H2 database driver to 1.4.196
- AP-10717: Expression Engine: UI Improvements
- AP-10673: Consistent background colours for JavaScript scatter and line plot
- AP-10644: JS views: new aspect ratio setting "square" (1:1)
- AP-10636: H2O Integration: H2O Scorer nodes should push scores as flow variables
- AP-10577: Python: Allow Python environment selection via flow variable
- AP-10529: Container Input (Table) with optional table spec in JSON input
- AP-10489: Optimize Parquet Dictionary Encoding
- AP-10451: R Snippet Dialog: Allow eval current line when nothing selected
- AP-10425: Move Annotations to Back/Front context menu entries
- AP-10406: Turkish stopword list
- AP-10397: FileStoreCell: Support for managing multiple FileStores
- AP-10379: IO Improvements for Keras and Tensorflow
- AP-10334: Container Input (Table) with new option to expose example table to the API definition based on the input table
- AP-10292: One To Many node throws misleading error when the domain of the input table contains more than 60 distinct values
- AP-10268: Python: Jupyter notebook support for Python nodes
- AP-10262: Improved error message for Call Workflow (Table based)
- AP-10221: Layout changes to support nested layouts (from nested wrapped metanodes)
- AP-10220: Nested Wrapped Metanodes behave the same as WizardNodes
- AP-10218: JavaScript scorer improvements
- AP-10136: Call Workflow (Table Based) improve invocation naming to reflect actual behavior
- AP-10075: Upgrade Tika version to latest version
- AP-10068: Column Expressions: Allow Conversion of DataCells
- AP-10037: Crosstab: improve default settings for Cross tabulation view
- AP-10034: Shapiro Wilk Normality Test: support multiple columns
- AP-10027: Wilcoxon Signed-Rank: compute median per column
- AP-10026: Statistics Nodes: show p-values with Full Precision by default
- AP-9845: Hyperparameter Optimization: Random Search
- AP-9814: Java Snippet to provide meaningful error message in log file
- AP-9799: Discard after execution option for Call Workflow (Table Based) node
- AP-9739: Turkish Tokenizer
- AP-9718: Make Cross Joiner Streamable
- AP-9610: Make Color Manager able to dynamically update possible values when using Palettes
- AP-9603: H2O Nodes - Better progress reporting when run with Sparkling water context.
- AP-9566: Column Expressions: New Functions to read / manipulate / create cell collections
- AP-9564: Workflow Editor: yellow message bar ("you are editing a remote workflow") always to be pinned to the top of the current view port
- AP-9526: Cell Splitter with option to guess number of columns based on first rows only
- AP-9525: Line Reader with option to read first row as column header
- AP-9524: Cell Splitter: "Split header column"
- AP-9492: Settings option to issue a warning when a node is replaced/inserted
- AP-9419: Python: Update Apache Arrow version to 0.10.0
- AP-9380: Upgrade Excel processing library (Apache POI) to version 3.17 (fixes issues reported in forum)
- AP-9357: Send Credentials via Call Workflow (Table Based)
- AP-9273: Retain order of nodes in wrapped metanode layout editor
- AP-9250: Text Mining: Add possibility to use custom model for OpenNLP tagger
- AP-9082: Keyboard actions to zoom in/out in workflow editor
- AP-8997: Lower severity level of "possible deadlock" message detector
- AP-8853: Tableau: Append to TDE file
- AP-8638: JS Views bidirectional communication support - core changes
- AP-8593: More intuitive paradigm to edit and move annotations
- AP-8470: Python: Allow cancelling data transfer
- AP-8266: Speed up of parameter optimization loops
- AP-7939: Re-enable the "Fullscreen" button in javascript views that aren't running in the WebPortal
- AP-7561: (API): ColumnRearranger to filter out unused columns when reading reference table (significant speed-up for large workflows when using Parquet table format)
- AP-7248: "Timer Info" node also recording statistics of nodes contained in meta nodes (non-wrapped)
- AP-6869: New Date/Time support in Tableau nodes (Hyper only)
- AP-6139: JavaScript Pie Chart: support selection
- AP-5352: Cell Splitter with option to replace input column
- AP-5233: Pivoting node needs column naming options
- BD-808: (Big Data Extensions): PySpark support for Local Big Data Environment
- BD-803: (Big Data Extensions): Upgrade "Create Local Big Data Environment" node to Spark 2.4
- BD-802: (Big Data Extensions): Add Spark 2.4 to KNIME Extension for Apache Spark
- BD-799: (Big Data Extensions): Support KNIME Workflow Executor for Apache Spark with local big data environments
- BD-762: (Big Data Extensions): Add type handling support to Parquet reader and writer
- BD-761: (Big Data Extensions): Add type handling support to ORC reader and writer
- BD-548: (Big Data Extensions): Support cancelation of wrapped metanodes with KNIME Workflow Executor for Apache Spark jobmanager
- BD-356: (Big Data Extensions): Selectable job manager for KNIME Workflow Executor for Apache Spark
- BD-318: (Big Data Extensions): Configurable persistence level of DataFrame created by KNIME Workflow Executor for Apache Spark
- BD-317: (Big Data Extensions): KNIME on Spark: Add support for ingoing flow variables
- BD-316: (Big Data Extensions): Support preference import from KNIME Server into KNIME Workflow Executor for Apache Spark
- BD-128: (Big Data Extensions): Enable arbitrary input ports with KNIME Workflow Executor for Apache Spark jobmanager
Bug Fixes:
- AP-10873: Bar Chart legend overlaps bins under specific requirements
- AP-10851: JavaScripts nodes blocking KNIME Server (on Oracle Linux)
- AP-10834: JavaScript Barchart does cut off visualisation early when stacking is used
- AP-10800: CTRL SHIFT ENTER does not execute node and open outport view
- AP-10555: Create Directory node fails
- AP-10884: XGBoost Locale Problems
- AP-10876: Resource leak in SftpURLConnection
- AP-10874: Improve EXAMPLES Server login failure message
- AP-10853: TagCloud: one word is colored wrong
- AP-10845: Connection loss when replacing output node in meta-node
- AP-10839: Default password in credentials input node does not work with REST interface
- AP-10823: Configuration dialogs of various learner nodes do not update twin list after new target is selected
- AP-10802: Parameter Optimization Loop Start outputs rounding errors
- AP-10792: Decision Tree Predictor fails if the input contains a column with the prediction column name
- AP-10773: Card View: Title & menu take up too much space
- AP-10756: Random Search: Make use of stepSize
- AP-10741: Wrapped-Metanode: Input / Output nodes can be deleted after changing port configuration (but shouldn't)
- AP-10727: Improve error messages for Download node when file does exist in Google Drive
- AP-10726: Can't resize annotations in zoomed in/out workflows
- AP-10710: No varargs possible for string-functions in Expression Engine
- AP-10693: Some views do not have full screen buttons
- AP-10659: JS Parallel Coordinates plot doesn't respond to selection in a composite view when the RowIds have complex format
- AP-10654: Random Forest Learner (dialog): When changing target, the old value goes into the 'exclude' list -- should be included
- AP-10653: Container Input (Table) silently accepts invalid input (e.g. "blah" in an int column is represented as missing)
- AP-10638: "Decision Tree to Ruleset" node fails on boolean splits
- AP-10569: Column Expressions node not parsing argument correctly when starting with (
- AP-10547: H2O: Column type lost if passed through node
- AP-10543: Math Formula (Multi Column) with RegEx throws error
- AP-10531: Call Workflow node fails to send URI columns to Container IO nodes
- AP-10528: Strange behaviour of String Manipulation (Flow Variable) node inside of Counting Loop
- AP-10511: JS Parallel Coordinates: Selection lost when resizing window
- AP-10447: Wrapped Metanodes invert order of flow variables
- AP-10434: Column Splitter shows warning if not-enforced columns are not available
- AP-10419: Range Slider Filter Definition node: Rounding min/max range issue
- AP-10409: Typo in "Reset not allowed" dialog.
- AP-10354: Parallel Coordinates plot cuts long labels on the first axis
- AP-10331: DL Keras: Sequential to Model converting different in Keras 2.2.0
- AP-10280: MacOS app bundle contains multiple instances of Chromium, resulting in excessive disk usage
- AP-10279: Integer Overflow in BufferedDataTable's incrementing IDs
- AP-10254: Table View does not respond to filters on excluded columns
- AP-10144: Table Validator does not convert integer to doubles
- AP-10122: Configuration dialog doesn't show when KNIME is not the active window on macOS
- AP-10109: Create Collection Column: typo in the node description
- AP-9943: Javascript Box Plot cuts long label names on y axis.
- AP-9871: Metanodes in Wrapped Metanode will fail without reason under detailed circumstances.
- AP-9775: Automatically connect DL nodes in the workflow by double clicking the node in Node Repository
- AP-9669: DL Keras: Node Generation Spinner Widget Bugs
- AP-9628: Workflow Editor: Replacing a node destroys customized link geometries
- AP-9581: Node (Un)Linking Hot Buttons do not work with Metanode boundaries
- AP-9455: H2O Integration: Stratified Partitioning problem (missing token, incorrect number of arguments, ...)
- AP-9247: Create FileStore Column in LoopEnd causes test workflows to fail
- AP-9167: String Manipulation (Variable) displays incorrect default value for "Replace Variable" when one has been set already.
- AP-8973: Unexpected display behavior when choosing filepath in quickform
- AP-8923: Tableau Server node description refers to Visual Basic 2013, but links to Visual C++ 2013
- AP-8711: H2O Cluster Assigner Duplicate column names
- AP-8126: Column aggregator needs to handle duplicate column names
- AP-8114: Excel Reader produces wrong output for Japanese characters
- AP-7592: File Reader fails on spaces in URL (should handle gracefully -- e.g. when using knime:// URLs)
- AP-5759: Weka 3.7 splash icons not shown
- AP-5653: JavaScript Box plot does not handle special doubles correctly (Nan and infinity)
- AP-5245: Naive Bayes Predictor: numerical problems with close-to-0-variance attributes + predicted probabilities slightly off according to PMML standard
- AP-4780: kNN node creates non-standard names for probability columns
- BD-806: (Big Data Extensions): Prevent IO in configure of FileFormatWriter/ReaderNodeModel
- BD-778: (Big Data Extensions): Spark SQL port causes unpersisting of cached DataFrames