KNIME Analytics Platform 3.6.2
Release date: Nov 6, 2018
Enhancements:
- AP-10242: Speed up execution of Bar Chart (JavaScript) and Pie/Donut Chart (JavaScript)
Bug Fixes:
- AP-10383: Context Menu dialog opens when applying changes in workflow annotations on Mac OS
- AP-10377: Download/Upload from List fails when downloading files
- AP-10342: Bit vectors have the same icon as booleans
- AP-10580: Streamed Wrapped Metanodes can't be properly canceled when in 'queued' state (leaves workflow in inconsistent state)
- AP-10574: Numeric Median Test: IndexOutOfBoundException
- AP-10540: JS Parallel coordinates plot should order categorical axes lexically
- AP-10496: H2O Integration: H2O MOJO Writer throws NPE when writing back to the server
- AP-10484: DB locking might freeze KNIME server
- AP-10468: The output ports are swapped in H2O Cross Validation Loop Start node
- AP-10456: File Reader hangs when using Quick Scan for large files (no progress bar but background proc ongoing)
- AP-10436: Call Remote Workflow node uses too low network timeout
- AP-10431: Column Type Auto Cast node: Inconsistent behavior for empty string handling
- AP-10426: S3 Delete Files fails without checking the option to abort on fail
- AP-10382: Text Mining: DmlDocumentParser erase last parsed term due to a referencing issue
- AP-10370: Text Mining: Document Data Assigner fails if file store chunk size is set to a low value in preferences
- AP-10363: Window Loop Start fails when run concurrently
- AP-10341: Not supported style rules in SVGs might lead to null pointer exception in Batik 1.7
- AP-10332: Bar Chart (JavaScript), new option to display maximum values
- AP-10295: Scorer Node Description: Link "accuracy statistics" is not correct
- AP-10263: Decision Tree Predictor fails if too many class values on append probabilities.
- AP-10219: JavaScript scorer: disabling "Display confusion matrix totals as rates" disables the totals completely
- AP-10216: Math Formula (Multi Column) fails on calculating statistics for non "CURRENT_COLUMNS"
- AP-10197: 100 % accuracy in the view of the Scorer (JavaScript) node even though some samples are not classified correctly
- AP-9811: Table Reader fails with "No entry data.bin in file" on non table files
- AP-9255: Wrapped Metanode - Node Usage Configuration Tab is missing Dialog Pane
- AP-7931: Python: Correct support for float32 in python bridge
- AP-7761: R_HOME value cause minor UI crash with invalid value
- AP-6190: X-Partitioner is missing description for Linear Sampling
KNIME Analytics Platform 3.6.1
Release date: Sep 3, 2018
Enhancements:
- AP-10071: Expression Engine: Change Math Functions to use a JavaMethodProvider
- AP-9797: Call Workflow (Table Based) improve user error messages
- AP-9048: DB Connector node: Strip leading and trailing white spaces from jdbc url
Bug Fixes:
- AP-10240: OSM Map View: MapQuest tiles broken
- AP-10192: JS Table View: "Subscribe to filter" menu option has no effect
- AP-10039: Eclipse help doesn't work any more due to wrong bundle resolution
- AP-10014: JSON to Table and JSON Path nodes throw errors when keys contain a comma
- AP-9948: JavaScript bar chart NP when plot horizontal bars is selected with image generation
- AP-9888: Iframe wrong resizing different quickform views to a minimum of 150px
- AP-9887: Label of range slider has now huge black font as a title
- AP-9885: JS Bar Chart losing interactivity on mouse-over when in wrapped metanode view
- AP-9117: Fix OSM preference page
- AP-10294: Don't show individual interactive views in context menu of a wrapped metanode
- AP-10284: Opening exported workflow without data may cause NPE in JS views and quickform nodes
- AP-10243: H2O Integration: Wrong node types/colors
- AP-10232: Java Snippet: Missing Converter Cache causes exception when creating file store cells on output
- AP-10195: IllegalStateException during execution of Chunk Loop
- AP-10184: Missing Value node sometimes cannot be loaded
- AP-10145: H2O Integration: Predictor nodes report missing columns if column is of wrong type
- AP-10135: Strip function description in String Manipulation and Column Expression node
- AP-10127: DL Keras: Forbid to output the input-layer in executor
- AP-10117: Multiple dragged nodes delete single node
- AP-10115: Fix label position of 'load'-button in node monitor
- AP-10105: Selected back end 'Keras (TensorFlow)' is not available anymore
- AP-10103: Unhandled event loop exception when trying to insert a node between two connected nodes inside a meta node
- AP-10084: The WebPortal/Wrapped Meta Node option gets checked every time configuration of a JavaScript View node is changed
- AP-10081: Expression Engine: Strings get wrongfully recognized as identifiers
- AP-10070: Expression Engine: Function average does not skip missing values and NaN
- AP-10061: DL Keras: Failing when keras runs in parallel
- AP-10060: DL Keras: Unconfigured keras network learner throws NullPointerException
- AP-10055: DL Keras: Network Reader loses executed state if file not available and not copied
- AP-10048: PMML: Load (compiled) Model Content fails under Windows
- AP-10025: Document Vector Applier does not adjust columns order before creating a collection cell
- AP-10016: Problem restoring mountpoint state
- AP-10010: DeepLearning4J: Better error message on OutOfMemory Error
- AP-9968: Call Workflow (Table Based) does not close called workflows on dispose
- AP-9966: Excel Reader Timeouts too small (causes some internal test cases to sporadically fail)
- AP-9942: DL Keras: Custom loss fails for Model API
- AP-9933: Missing Value node slows down workflow load
- AP-9906: Call Workflow (Table Based) outputs non informative NPE error message from Non-Native data value
- AP-9884: Expression Engine: Fix creation of LocalDateTime
- AP-9875: Wrapped Metanode Selection Interactivity lost after closing and re-opening KNIME
- AP-9857: BarChart Tooltip display inconvenience
- AP-9843: DL Keras: Regularizers, Initializers, Constraints are not validated
- AP-9675: Call Workflow (Table Based) can not be configured when in temporary state
- AP-9476: HTTP/HTTPS Connection nodes ignore KNIME's proxy settings (configured in 'Preferences')
- AP-9454: KNIME Protocol not properly resolved for Temp Workflows when workspace contains space character
- AP-9439: Text Mining: French Stanford POS model uses another tag set than FTB
- AP-9410: Text Mining: StanfordNLP PTBTokenizer does not work in some cases
- AP-9149: Workflows fail when Python nodes run in parallel
- AP-8850: Bar Chart: shows last Missing Value Warning, even if those are resolved.
- AP-8829: JavaScript Bar Chart only shows one decimal digit in y-axis label
- AP-8612: Python: Distinguish between SocketException causes during kernel startup
- AP-7306: JavaScript Bar Chart truncates labels
- AP-6424: Kruskal Wallis test - Contradiction between R and KNIME
- BD-775: (Big Data Extensions): Livy Job Controller changes order of files in Job with Files Implementations
- BD-772: (Big Data Extensions): Create Local Big Data Environment fails when hostname contains invalid characters
- BD-771: (Big Data Extensions): KNIME-on-Spark jar collection does not work in KNIME 3.6
- BD-726: (Big Data Extensions): Parquet writer node allows wrong input for filename
- BD-689: (Big Data Extensions): Spark Reader/Writer nodes do not support spaces in file path
- BD-624: (Big Data Extensions): Create Local Big Data Environment: Create-Destroy-Create fails when Hive data folder is set
- BD-607: (Big Data Extensions): Deactivating impersonation on KNIME Server with Simba drivers results in "GSS initiate failed"
KNIME Analytics Platform 3.6.0
(see highlight summary)
Release date: July 11, 2018
New Nodes:
- AP-9873: Kafka Integration (Labs; several nodes to produce and consume messages from Apache Kafka)
- AP-9869: Deep Learning Tensorflow Integration (nodes to read/write and run TensorFlow deep learning models)
- AP-9867: Deep Learning Keras Integration (major enhancements and 50+ new nodes to design deep learning networks)
- AP-9863: H2O Sparkling Water Integration (Run any H2O KNIME Workflow on Apache Spark using "H2O Sparkling Water")
- AP-9862: New Database Integration (25+ node nodes; enhanced capabilities, e.g. connection pooling, JDBC dialect support, feature richt SQL editor, rich type support, etc.)
- AP-9472: Call Workflow (Table Based)
- AP-9292: Service Variable Input (a node to simplify assembling workflows that are called by other workflows or via REST interface)
- AP-9104: Service Table Input (a node to simplify assembling workflows that are called by other workflows or via REST interface)
- AP-8996: Constant Value Column Filter
- AP-8864: Numeric Outliers (Apply)
- AP-8745: Column Expressions (apply advanced expressions to modify or append a number of columns)
- AP-8732: Git Client Integration (a number of nodes to read information from Git repositories)
- AP-8676: Dictionary Tagger (Multi Column)
- AP-7707: Text Mining: New Stop Word Filter (revised)
- AP-7247: Text Mining: Term Neighborhood Extractor
- AP-6411: Javascript Scorer Node
- AP-5355: Numeric Outliers
- BD-682: (Big Data Extensions): Spark Association Rule (Apply)
- BD-681: (Big Data Extensions): Spark Frequent Item Sets
- BD-596: (Big Data Extensions): Create Spark Context (Livy)
- BD-577: (Big Data Extensions): ORC File Writer
- BD-576: (Big Data Extensions): Parquet File Writer
- BD-572: (Big Data Extensions): ORC File Reader
- BD-571: (Big Data Extensions): Parquet File Reader
- BD-547: (Big Data Extensions): Spark Missing Value (Apply)
- BD-534: (Big Data Extensions): Spark Association Rule Learner
- BD-532: (Big Data Extensions): Spark Pivot
Enhancements:
- AP-9219: Upgrade underlying Eclipse framework to version 4.7.3
- AP-9515: Text Mining: Update node description of POS Tagger and OpenNLP NE Tagger
- AP-9510: Flow Variable of type date can't be used as flow variable input
- AP-9496: Adding Padding Method to String Manipulation and Expression Engine
- AP-9480: Consistency of messages for temporary workflow copies opened in workbench
- AP-9421: Update JSON Path library to 2.4.0
- AP-9420: KNIME SVG Support Extension as part of Standard Build
- AP-9251: Bundled Chromium for Mac and Linux
- AP-9223: KNIME protocol relative paths should work in when editing server workflows
- AP-9221: Workflow Difference to also inspect contents of Wrapped Metanodes
- AP-9193: Include global and user generated CSS in all JavaScript views
- AP-9139: H2O Integration: Improve node description for scorer and predictor nodes
- AP-9127: Node Generation: Implement prototype for new annotations
- AP-9032: Excel style selection of multiple cells and copying their values in Table View (JavaScript)
- AP-8989: Improved persistence of server mountpoints in Analytics Platform
- AP-8977: Speedup Logistic Regression Predictor, especially for 'wide' data/fingerprints (tens of hundreds of attributes)
- AP-8976: DL Framework: Rename nodes
- AP-8965: R Integration: Improve detection of R Installation
- AP-8925: Update Infocom/ChemAxon extension to 3.6.0.v0901
- AP-8916: DL Framework: Make environment setup cancelable
- AP-8892: Recursive Loop End(s) with Option to choose FlowVariable in Dialog
- AP-8840: DL Framework: Integer collection to one-hot vector converter
- AP-8823: Amazon S3 nodes to support Server-Side Encryption
- AP-8748: Column-based storage: Support more than just native types
- AP-8744: Improve performance of DataCell#getType()
- AP-8731: DataTableDomainCreator: Improve performance
- AP-8723: Make Ungroup-node streamable
- AP-8704: New table storage format for KNIME tables based on Apache ORC (labs)
- AP-8683: Column Filter ("Twin List") with improved filter option
- AP-8636: (API) Credentials to be available during loadInternals()
- AP-8614: URI Ports selectable in Metanode's Add Port dialog
- AP-8613: Support the extended character limit for Tweets
- AP-8598: Text Mining: Remove document cell type selection from preference page
- AP-8592: Placeholder if no results are found in node repository
- AP-8568: Switch from PhantomJS to headless Chromium for image generation in JavaScript views
- AP-8555: Move individual views from context menu of wrapped metanodes into submenu
- AP-8544: JavaScript Table Editor: support keyboard navigation
- AP-8543: JavaScript Table Editor: support pasting multiple values
- AP-8494: DL Framework: Auto-mapping of dimensions for multi-dim inputs
- AP-8451: Add better default color palette option to Color Manager node
- AP-8447: File Upload supporting S3 buckets with permissions managed with IAM policies
- AP-8432: Python Framework: Pure message driven communication
- AP-8360: Add in-memory credentials to Google Sheets Interactive Service Provider
- AP-8351: DL Framework: Keep column selection for networks with same input / outputs
- AP-8275: Dialog Component date to set time to 0 on empty all fields
- AP-8268: DeepLearning4J: Preference page options are not applied
- AP-8036: DL Framework: Support for not or partially defined input shapes
- AP-7884: H2O Integration: Inform user about column matching problems in H2O Predictor nodes
- AP-7682: "FTP Connection" node with to support proxies
- AP-7560: Column-based storage: New Parquet table format
- AP-7558: (API) Extension point for table serializers (Parquet, ORC, Standard-KNIME)
- AP-7532: Replace variables in preferences files before application
- AP-7531: Write additional profile files to workspace directory
- AP-7530: Create plug-in customization file from preference profiles
- AP-7529: Add extension point for preference profile providers
- AP-7514: H2O Integration: Variable Importance Measurement Output DRF / GBM
- AP-7513: H2O Integration: Partitioner Options
- AP-7001: Java Snippet: New dialog tab to add installed bundles to code classpath
- AP-5673: Allow connecting and unconnecting nodes using keyboard shortcuts
- AP-5238: Node Replacement/Insertion to also work when moving nodes
- AP-5062: Better Zooming in workflow editor: as low as 25% zoom level and ctrl+mousescroll to zoom in/out
- AP-3136: Update Apache Batik to 1.7
- BD-685: (Big Data Extensions): Add remote file handling input port to Create Spark Context (Livy) node
- BD-666: (Big Data Extensions): Add Spark 2.3 to KNIME Extension for Apache Spark
- BD-659: (Big Data Extensions): Update hadoop-client library from 2.7.4 to 2.7.6
- BD-635: (Big Data Extensions): Reduce Spark Exception error messages length
- BD-629: (Big Data Extensions): Big data file format support for Spark to Hive (Impala)
- BD-611: (Big Data Extensions): Use new Spark DataFrame API in preprocessing nodes (partition, sampling, sort, rename column and join)
- BD-609: (Big Data Extensions): New seed parameter in Spark Collaborative Filtering Learner node
- BD-507: (Big Data Extensions): New seed parameter in Spark k-Means node
- BD-450: (Big Data Extensions): Revise Spark PCA node
- BD-217: (Big Data Extensions): Add option to replace original columns to Spark Transformations Applier
- BD-47: (Big Data Extensions): Improve append to existing Hive/Impala table with Hive/Impala Loader
Bug Fixes:
- AP-9672: Column/value selection quickforms broken in Firefox
- AP-9494: Table Writer has path from history set on create
- AP-9762: Table to H2O does not filter out String columns with only missing values
- AP-9735: Text Mining: Dictionary Tagger does not find repetitive patterns and does not tag all entity occurences
- AP-9651: Writer nodes should handle cancellation gracefully
- AP-9647: Settings are not stored when restarting after prompt
- AP-9643: H2O Add Limit To Number of rows being cached.
- AP-9642: JavaScript scorer node: Non-informative error on numerical columns
- AP-9618: Learning monitor shows nothing when "log scale" is selected in Accuracy mode
- AP-9598: Curved connection created between nodes outside the viewport doesn't show up (fully)
- AP-9594: DL Framework: Selection of inputs via Pattern doesn't work
- AP-9593: Pie/Donut Chart not executed when frequency column value is 0
- AP-9588: Nullpointer in Value Selection Quickform
- AP-9585: File access problem while saving workflow
- AP-9556: Line Plot (JavaScript) fails to create SVG, if too many Y axis columns
- AP-9554: Problem opening two JavaScript-Views at the same time
- AP-9534: Possible message during update: "An exception occurred while executing a runnable."
- AP-9518: Determining MAC addresses on Ubuntu 18.04 fails due to missing ifconfig command
- AP-9516: DL Keras: DLKerasTuple is not grayed out if optional
- AP-9497: DL TensorFlow Reader: possible race condition
- AP-9483: H2O Integration: Table to Frame casts to TypeCell instead of TypeValue
- AP-9481: Standard BIRT template out of date
- AP-9449: Tree Ensembles: Serialization bug for nominal columns with many possible values
- AP-9447: H2O Integration: Random Forest Learner crashs unexpected for very small data sets
- AP-9404: 'Download / Upload from List' fail to create parent folders
- AP-9397: H2O Integration: Target column present in column selection when dialog opened
- AP-9374: Column Auto Type Cast fails if converted column is all missing values
- AP-9368: Deep Learning Learning Monitor view too large vertically
- AP-9356: Call Local Workflow occasionally does not update Path in configuration dialog
- AP-9283: UI Width of element in Row Filter too small
- AP-9261: Zipped files in workflow will not exported when 'Reset workflow(s) before export' enabled
- AP-9228: File chooser component does not reset after 'cancel' in dialog is clicked
- AP-9215: MacOS: Clear Console Crash
- AP-9212: Wrapped metanode templates don't receive username when newly created.
- AP-9203: Null pointer exception in Random Forest
- AP-9201: Quick Node Insertion hotkey (ctrl-space) does not select text already present in search box
- AP-9151: Null Pointer in Data Explorer
- AP-9119: R Integration: Errors during printing a value are not properly escaped
- AP-9118: Unzip Files: Null pointer exception when setting flow variable for output directory
- AP-9111: DL Framework: Invalid converter when input changed
- AP-9109: Filename persists in dialog despite user hitting cancel (Windows only)
- AP-9089: DL Framework: Investigate converter matching issues
- AP-9055: H2O Integration: Fix functionality of H2O Binomial Scorer
- AP-9053: Red X stays on node if node replacement is aborted with ESC
- AP-9052: Scatter Matrix node throws nullpointer for emtpy string column
- AP-9009: Java Snippet Dialog does not show missing jar files
- AP-8987: Workflow Coach locks KNIME when configured without Internet
- AP-8985: (Workflow) Annotation text possibly truncated when workflow zoomed > 150%
- AP-8724: Inconsistent Names of Prediction Columns in TreeEnsembleLearner (space missing)
- AP-8685: Abnormal memory consumption in ExplorerFileStoreTransfer
- AP-8682: Improve NameFilter save/load filter configuration performance
- AP-8680: Table Creator: Inconsistent right-click behaviour on column header
- AP-8677: DL Framework: Dialog cannot be opened: failed to load config
- AP-8665: XML documents are not thread-safe
- AP-8607: Workspace Launcher is called Eclipse Launcher
- AP-8586: Workflow Coach Preferences can be set even if not enabled
- AP-8577: H2O Integration: Column Selection does not save Wildcard/RegEx and Type Selection
- AP-8365: Workflow Coach View: Truncated text
- AP-8364: DataColumnSpecFilterPanel: Extremly slow for tables with many columns
- AP-8018: Dialog hot keys for Ok+execute is not reliably working
- AP-7835: DataColumnSpecFilterPanel: Changes of the input column type lead to inclusion of filtered out columns when enforce inclusion is selected
- AP-6886: Parallel chunk loop fails if in wrapped meta node
- AP-6821: Width of UI element in POST Request node dialog overly large
- AP-6558: R nodes don't provide full error message
- AP-6267: Unzip file node fails with NPE on invalid file
- AP-5670: Keyboard shortcuts not consistently working on MacOS
- AP-3134: SVG output in pie charts is broken when one Pie is >0.5
- BD-714: (Big Data Extensions): Remove confirmation dialog in "Destroy Spark Context" node
- BD-712: (Big Data Extensions): Spark Joiner fails if column names contains dots
- BD-699: (Big Data Extensions): Parquet to Spark does not allow to select file when Source is invalid
- BD-697: (Big Data Extensions): Spark PCA node fails to refresh the target dimension control after clicking the Flow Variables Tab
- BD-689: (Big Data Extensions): Spark Reader/Writer nodes do not support spaces in file path
- BD-667: (Big Data Extensions): Spark "mean" missing value handler should change column type to double
- BD-652: (Big Data Extensions): Spark Learner and Scorer nodes should handle booleans
- BD-642: (Big Data Extensions): Spark PCA/SVD/Correlation Matrix on Spark 2.x produce different specs in KNIME and Spark
- BD-634: (Big Data Extensions): Mass column renaming on Spark 2.x takes too long
- BD-627: (Big Data Extensions): Spark Transformation Applier produces broken DataFrame/RDD for type-changing transformations
- BD-612: (Big Data Extensions): Spark table spec contains possibly wrong metadata
- BD-393: (Big Data Extensions): Spark context state management prevents users from reusing RDDs