Changelog v2.10.0

Changes from v2.9.4 to v2.10.0 (see highlight summary)

Release Date: July 10, 2014

New Nodes:

  • New Node 5196: (Several nodes) JDBC Connector nodes for different database systems
  • New Node 5197: Database Table SELECT node
  • New Node 1968: Database SQL Execution node
  • New Node 5194: Database Sorter node
  • New Node 5195: Database GroupBy node
  • New Node 1675: Database Joiner: with two DB in-ports and one DB out-port
  • New Node 5219: Database Reader, Writer, Update, and Delete nodes should use the new database connection port + input table
  • New Node 5193: SQL Inject and SQL Extractor nodes to merge and extract SQL statements from a Database / JDBC connection
  • New Node 4643: String/XML to SVG (casts String or XML column into SVG)
  • New Node 4907: (Several nodes) binary and byte vectors fingerprint handlers (conversion to/from collection cells etc)
  • New Node 5263: (Several nodes) Parameterized distance functions: numeric, string, fingerprints, aggregated, snippet, matrix, mahalanobis
  • New Node 5123: (Several nodes) Google API (Analytics for now)
  • New Node 5165: (Several nodes) Twitter API
  • New Node 4309: Data Statistics w/ skewness, distribution graph etc.
  • New Node 5336: Text Mining: Topic extractor node
  • New Node 5215: Network mining: New Hierarchy Extractor node
  • New Node 4642: Read Images for basic types (SVG; PNG) -- replaces Read PNG Images node
  • New Node 4982: XML to PMML node
  • New Node 5146: Time Series: Moving Aggregation node
  • New Node 5199: Time Series: Missing Value
  • New Node 5329: DateTime Shift: add/subtract to/from date and time (add 1 day, get previous month, ...)
  • New Node 4534: (Data Generation) General number generation
  • New Node 5220: (Data Generation) Random Boolean Assigner

Enhancements:

  • Enh 4944: New intro page upon each KNIME startup
  • Enh 1600: Auto-saving of workflows
  • Enh 4988: Update to Java 7 Update 60
  • Enh 4899: R integration (labs) to work with R v3.x by default (it's currently linked to R v2.x)
  • Enh 5273: New R (Labs) nodes moved top-level (replacing old R nodes)
  • Enh 5075: New meta nodes for time series analysis ("Seasonality Correction", "Time-Series Auto-Prediction Training", "Time-Series Auto-Prediction Predictor")
  • Enh 4601: Column Filter dialog to show names of deliberately included/excluded columns even if no longer available in input
  • Enh 3588: New grouped node concept: Subnodes (Metanodes need "isolation" mode.) (Labs)
  • Enh 4602: Rewrite of existing column rename node (better handling of many columns and long column names)
  • Enh 4866: Network mining: Subgraph Extractor option to include only reachable objects in a directed network
  • Enh 5309: Make database timeout configurable via preference page
  • Enh 4056: Radio String Quickform node should have multiple display modes.
  • Enh 5375: Table Reader node with reasonable progress and ability to Cancel; also new control to limit number of rows
  • Enh 4862: (API) NodeLogger with additional methods to compose log messages (e.g. info(String template, Object... args))
  • Enh 5152: K-Medoids node to have view with "Finish" button + progress/cost-reduction table (early abort)
  • Enh 2167: Database Connection nodes must quote column with space
  • Enh 4865: GroupBy, Column Aggregator and Pivoting node to support escape characters for string concatenation ('\n', '\t')
  • Enh 4846: 1.Normalizer node with additional options for column filtering (enforce include/exclude, name pattern matching)
  • Enh 4879: Enhancements to "Cross Joiner" node (faster execution, more options)
  • Enh 4544: Add node relative URLs for knime-protocol (e.g. knime://knime.node/drop/data.txt)
  • Enh 4909: Network mining: Network Viewer with option to hide arrows of directed edges
  • Enh 3650: Add a few pre-configured R meta nodes including Grouped StatterPlot and Stacked BarChart
  • Enh 4736: Autobinner should also equally distribute count and use integers better
  • Enh 3615: Timeout of Read (PNG) Image node should be adaptable by the user
  • Enh 4954: KNIME Explorer support for .sqlite files.
  • Enh 4863: (API) New (ColumnSelectionSearchableListPanel) utility panel to configure multiple columns (used in EditNominalDomain/Rename/MissingValue/...)
  • Enh 4973: Make node message test for test workflows optional
  • Enh 4971: Output stacktraces for timed out test workflows
  • Enh 4980: Network mining: New extension points for Network Generator node modules
  • Enh 4985: Ship KNIME on MacOS X with a recent JRE
  • Enh 5207: Save plug-in version for each node in workflow
  • Enh 4917: "Tree Ensemble" Learner with different default values in the configuration (Random Forest defaults with 100 trees)
  • Enh 5013: "Tree Ensemble" (Random Forest) learner node to be more memory friendly
  • Enh 5019: Automatically close testflows when test was successful
  • Enh 5018: Execute workflow tests on multiple workflows at once
  • Enh 5016: Hilite rows during workflow tests
  • Enh 4655: Adapt SOTA nodes to new predictor guidelines
  • Enh 5125: Define JSON interchange format for JavaScript grid layout
  • Enh 1532: (API) Added utility methods for bit vector handling - increased speedup
  • Enh 5114: (API) Node Description xml file should allow a description tag within a 'tab' (better layout)
  • Enh 5050: Rule Engine nodes should provide table specs on configure.
  • Enh 5087: Workflow Editor should allow scrolling/panning with middle mouse button
  • Enh 5144: Date (String) Input should use same data format as String to Date Time
  • Enh 5147: Show information on missing browser integration
  • Enh 5262: New extension point for distance function contributions (allows parameterized distance functions)
  • Enh 4582: Polynomial regression learner should output coefficients and p-values
  • Enh 5283: Add timeout option to filehandling connection nodes
  • Enh 4986: Image To Table node with "Column Name" option (and flow variable support)
  • Enh 5264: Distance based nodes support parameterized distance functions (similarity search, k-medoids, hier. clustering, ...)
  • Enh 5037: Use plug-in for org.apache.commons.lang3 (version 3.2.1) instead of bundled Jar (now used in logistic regression, e.g.)
  • Enh 5110: R (Interactive): Nodes with table output should have checkbox to represent extremes (NaN & Infinity) as Missing Cells or as-is
  • Enh 5035: XLS Writer - Options to auto fit column width and to set print options (landscape/portrait, paper size)
  • Enh 5261: Node Monitor should show saved plug-in version
  • Enh 5191: Adaption of PMML Predictor and PMML Ensemble Predictor to use the new predictor implementations
  • Enh 5237: Double-click on sub node should open dialog, not open the editor. Editor can be opened via Ctrl+Double-Click
  • Enh 5129: BatchExecutor should name the workflow after the "-destfile" argument (currently always "BatchExecutorOutput")
  • Enh 4603: Upgrade SAS DSREAD Version 2.5, 2014-01-22
  • Enh 5172: Polynomial Regression Learner should have auto-guessing similar to Linear Regression Learner
  • Enh 5082: Use new column filter panel in One2Many/Many2One nodes.
  • Enh 5163: Threshold for number of possible values to store should be adjustable (currently hardcoded to 60 - set by -Dknime.domain.valuecount=xy)
  • Enh 3353: More flexibility for the Date Input Quickform node
  • Enh 4974: Network mining: Node Neighbor Extractor with new filter empty neighbors option
  • Enh 4824: Row Filter and Row Splitter should support deep filtering of collection values
  • Enh 3675: Value filter quickform node should provide all columns from the table.
  • Enh 3351: Value Selector QF supporting variable types
  • Enh 3674: Ability to lock and hide column selection in value selection quickform node
  • Enh 4055: Value Selection Quickform node should have multiple display modes.
  • Enh 4818: Column Filter: Should have preview twin list for type filter
  • Enh 5124: Layout option for new Javascript views
  • Enh 4926: ROC node should have option to sample datapoints for large datasets
  • Enh 5173: Polynomial Regression Learner should use new columns selection panel
  • Enh 5177: Naive Bayes Learner & Predictor to support PMML (requires PMML 4.2)
  • Enh 5168: Add "CML" type to list of "Molecule Type Cast" node
  • Enh 3541: Better widget for Quickform node in workflow execution wizard
  • Enh 3636: Default value of quickform nodes can't be controlled using variables
  • Enh 5171: Rule Engine dialog should allow escaping in regular expressions
  • Enh 4841: New Quickform Nodes should allow input validation (e.g. number range checking)
  • Enh 5176: Upgrade to PMML 4.2
  • Enh 5190: JPMML Upgrade to 1.1.3
  • Enh 5169: Fingerprint Bayes nodes to support multi-label problems
  • Enh 5116: "Create Collection Column" node should also work with empty column list (output is empty collection column then)
  • Enh 5192: New Database JDBC Port
  • Enh 5092: Support for tracks and routes in OpenStreeMap nodes
  • Enh 5211: Pipeline Pilot node to support "continueJobAfterTimeout" field
  • Enh 5093: OpenStreeMap View: Support hilite on a selection of map markers
  • Enh 4910: Network mining: New tree layouts for Network Viewer
  • Enh 5310: Option for disabling metadata retrieval during configure of database nodes
  • Enh 5227: (API) FileStorePortObject needs to support collection of FileStores (currently only one)
  • Enh 5022: Logistic Regression Learner improvements (better progress, cancellation, automatic filtering of constant columns, error messages)

Bug Fixes:

  • Bug 5118: Old R nodes may take up foreign workspaces
  • Bug 5099: Rule2DDrawingPane re-fires hilite events
  • Bug 5267: Linear Regression Learner uses different type filters in dialog and model
  • Bug 4858: Random Number Assigner dialog has unused option "bounds" and needs more description
  • Bug 5319: Text Mining: Sentiment tag set is missing
  • Bug 4857: Random Number Assigner throws NPE in dialog
  • Bug 4922: Adapt Weka predictors to new predictor guidelines
  • Bug 5239: Incomprehensibly Active Branch Inverter node help text
  • Bug 5238: DB Writer should add field names to INSERT statement
  • Bug 5111: FileStore objects cannot be created in LoopEnd (only be referenced from loop body)
  • Bug 5100: ClusterNodeView re-fires hilite events
  • Bug 4688: Table To Image node does not recognize the image type in the image spec (problems with SVG columns)
  • Bug 4693: Wrong #equals() implementation in SVGCells
  • Bug 5098: ScatterPlotter re-fires hilite events
  • Bug 4812: KNIME GUI: Workflow reset takes too long
  • Bug 5351: Database: support for long cells (BIGINT) for Database Reader and Writer
  • Bug 5331: Constant Value Node accepts the String Type only on set flow-variable
  • Bug 3012: CancelAll tries to also cancel configured nodes (resulting in a WARN-message)
  • Bug 5103: Hilite events are not processed in order all the time
  • Bug 4851: Network mining: Network Viewer does not work on KNIME Server (headless exception)
  • Bug 4066: Database Reader creates different row ids depending on fetch size
  • Bug 5229: Wrong passwords for database connections may result in strange error messages
  • Bug 4933: All node dialog panes are wrapped in a scroll panel to avoid layouting issues on smallish screens
  • Bug 4940: Loop End Nodes collect too many domain values from loop iterations in memory (should use standard count thresholds)
  • Bug 5327: Decision Tree Predictor fails with unreasonable error when PMML (generated by R) contains boolean target
  • Bug 4975: New interactive Javascript views concept (in KNIME Labs, not meant to be used in production just yet)
  • Bug 4961: AggregationMethod combobox should display aggregation methods in alphabetical order
  • Bug 4993: Create Directory: Option to configure the flow variable name
  • Bug 4984: Dialogs cannot be opened on MacOS X with Java 7
  • Bug 5258: Workflow SVG Export chokes on umlauts
  • Bug 3037: Renaming MetaNodes doens't recalculate node bounds
  • Bug 5036: Can set temporary directory in KNIME preferences to read-only location (should choke instead)
  • Bug 5031: JSpinner keeps manual changes if a node dialog is canceled (while the spinner still has the focus) and opened again
  • Bug 5060: Deadlock while opening a node view when node is still executing
  • Bug 5339: WorkflowLogMessagesTest not able to deal with expected duplicate error messages
  • Bug 5083: String escaping in the configuration dialog of rule nodes is confusing
  • Bug 5337: Generic Web Service node to have control flag to handle errors (fail if any invocation fails, fail if all invocations fail, never fail)
  • Bug 5089: Linear Regression Learner generates missing/NaN values when there are missing values in the input and nominal column in the learning data
  • Bug 5090: The offset in Linear Regression Learner gets applied (in some ways) even if the option to use it is unchecked.
  • Bug 5122: Domain Calculator: Extract logic in a separate class so that other nodes can benefit from.
  • Bug 5304: String manipulation error in method: removeChars
  • Bug 4998: Unify linear and polynomial regression nodes
  • Bug 5179: Ignore Eclipse command line parameters in batch executor
  • Bug 5117: Time Generator node gives incorrect date when time stamp is not generated
  • Bug 5301: Constant Value Node does not save value setting of flow var
  • Bug 5272: Plotter views choke on integer columns with large domains
  • Bug 5274: Synchronizing explorer with editor may select wrong mount point
  • Bug 5257: Missing estimators and searchAlgorithms in Weka's BayesNet configure dialog
  • Bug 5077: Some methods of SparseBitVector /DenseBitVector do not work correctly (e.g. xor - rarely used)
  • Bug 5268: DataColumnSpecFilterPanel should use type filter of the config given in loadConfiguration()
  • Bug 5313: Text Mining: Typos in dialogs and node descriptions
  • Bug 4929: "Rename" node cannot be used with variables (configuration tree is dependent on actual column names)
  • Bug 5366: Naive Bayes Learner: support for auto-guessing target column
  • Bug 5359: Opening the report perspective after save as goes to the previous workflow
  • Bug 5091: Text Mining: Concurrency issue in DocumentFileStoreCellFactory
  • Bug 4474: No missing node place holder for nodes contributed via DynamicNodeFactory
  • Bug 4470: In-build Meta-nodes in KNIME use deprecated nodes and should be updated
  • Bug 4684: Upgrade of free ChemAxon/Marvin nodes to newer version (2.8.2.v0205) - InCHi renderer/converter support + new renderer
  • Bug 5221: BitVector support added to Item Set Finder (Borgelt) and Association Rule Learner (Borgelt)
  • Bug 5104: Remove duplicated code in linear regression and logistic regression for determining the domain values.
  • Bug 5058: SparseBitVectorCellFactory#and(DenseBitVectorCellFactory) computes wrong result (only when types were mixed)
  • Bug 5033: ConcurrentModificationException in HiliteHandler if events come too fast
  • Bug 5333: k-Means node does not update include and exclude column list while re-configuring
  • Bug 4465: Headless workflow execution shouldn't require SWT display
  • Bug 5289: Text Mining: Document Viewer Node ignores case of term in search
  • Bug 5348: PMML Predictor support for Naive Bayes models (PMML v4.2)
  • Bug 4577: One2Many fails with illegal argument exception
  • Bug 5183: Variable Loop End node: Allow selection of variables
  • Bug 4035: Parsing SQL or R code containing flow variable identifiers causes infinite loop if type identifier is missing
  • Bug 4683: Extract and centralize domain calculation code
  • Bug 5189: Use latest SWT libraries to solve browser issues under Linux
  • Bug 5149: Log File should open in Eclipse Editor
  • Bug 5216: Renderer list in table view does not show adapter contents
  • Bug 5217: Database writer should allow columns with spaces
  • Bug 4780: Decision Tree Learner does not set default childs and causes an error in the Decision Tree Predictor
  • Bug 5315: Text Mining: Wildcard Tagger has problems with '#' in sentences.
  • Bug 5228: MLP Predictor dialog shows wrong value for "append probabilities"