Changelog v2.4.0

Changes from v2.3.4 to v2.4.0

Release Date: July 07, 2011

Enhancements

  • A set of enhancements around Meta-Nodes: Collapse / Expand and rename
  • Meta-Node dialog: Exposes the settings of quickform nodes contained in the meta-node
  • Auto-Layout and horizontal alignment (experimental)
  • Fast node insertion (double-click on node inserts and connects it)
  • Looping enhancements: stepping, more responsive GUI, parallel loop execution (experimental, available from KNIME labs)
  • Enhanced PMML functionality, support of PMML 4.0 including processing For details see this paper that will be presented on the KDD2011 PMML Workshop.
  • Enh 2605: Actions for step, pause and resume loop execution
  • Enh 2190: Database nodes may fail for databases that don't support aliases/subquerys
  • Enh 2437: GroupBy Dialog and Description needs improvement
  • Enh 2597: Enable Expert mode per default
  • Enh 2721: GroupBy node with variable delimiter pattern for Concatenate option
  • Enh 2726: Excel must have option to read default (=1st) tab (needed for quickform exec)
  • Enh 2744: Stressing node needs second outport for following changes
  • Enh 2746: Stresser node needs option for row/cell probability (MDG)
  • Enh 2747: Improve startup time for batch execution
  • Enh 947: CSVReader/-Writer should handle gzipped files
  • Enh 2507: New aggregation method that creates list, sorted list and set cell without missing values
  • Enh 2522: General missing value handling option for AggregationMethods
  • Enh 2611: Hierarchical Cluster Assignment needs option to do clustering based on similarity threshold (rather than cluster count)
  • Enh 2636: MoSS node should output molecules for each fragment
  • Enh 2716: Rule Engine node: Allow value of selected column in consequence part
  • Enh 2724: Feature Elimination node to have filter by error threshold option
  • Enh 2728: SDFWriter and SDFInserter need option to omit missing values
  • Enh 2730: MolfileReader needs option to extract molecule names into column
  • Enh 2731: SDF reader improvements (multiple files, limit rows, ...)

New Nodes

  • New Node 2754: Cross tabulation (with useful statistics such as fisher test and chi-square)
  • New Node 2732: Equal Size Sampling (down sampling such that values in class column are equally distributed)
  • New Node 2743: Pivoting (more aggregation methods and multiple group and pivoting columns)
  • New Node 2744: A new set of 11 new XML processing nodes
  • New Node 2749: Radarplot appender node (star/spider plot)
  • New Node 2548: PNG image writer
  • New Node 2624: Column renamer based on a regular expression
  • New Node 2701: Column Header Insert to change column names according to a dictionary table"
  • New Node 2633: Binner categorizing according to rule table (2 inputs, 1 output)
  • New Node 2647: Column Merger to merge two columns, choosing the non-missing value for the output
  • New Node 2675: Create Temp Dir creates temporary directory and expose its path via flow variable
  • New Node 2698: Loop start node for iterating over columns
  • New Node 2700: Line Reader node that reads entire lines from a file
  • New Node 2755: Bayesian Learner & Predictor for (Chemical) Fingerprints
  • New Node 2759: Association rule learner based on a-priori/ext.Tool algorithm
  • (No ID): Nodes for Ensemble Learning (Boosting, Bagging)
  • (No ID): SAS7BDAT Reader (available from labs)

Bug fixes

  • Bug 2756: SVG export fails on certain workflows (colors) and kills KNIME
  • Bug 2739: Set Operator has problems with multiple values
  • Bug 2138: Hilite state not propagated through hilite translation
  • Bug 2187: Syntax error when writing table to an SQLite DB
  • Bug 2203: Concatenate node removed hilting during execution
  • Bug 2558: Export Wizard with empty source creates empty archive
  • Bug 2598: HiLiteTranslator should not reset hilite when new mapping (null) is applied
  • Bug 2625: Joiner: Wrong behavior when joining columns with missing values
  • Bug 2640: GroupBy: Concatenate aggregation should produce missing values if all values in a group are missing
  • Bug 2672: R binary feature name misleading
  • Bug 2673: Some plug-in names don't start with KNIME
  • Bug 2674: KNIMEConstants#BUILD_DATE simply wrong (java doc)
  • Bug 2678: Annotation automatically increases font size each time you edit it (only when zoomed)
  • Bug 2681: Database credentials not updated
  • Bug 2688: Duplicate SVM in PMML of SVM Learner for binary classification
  • Bug 2690: Java Snippet Row Filter/Splitter do not update ROWINDEX field for 1st row
  • Bug 2691: Decision Tree View number formatting improvements
  • Bug 2692: Decision Tree View in Learner and Predictor uses different number formatting
  • Bug 2693: Decision Tree View - histogram normalization on #trainings pattern
  • Bug 2694: Auto-Binner should use independent internal NumberFormat
  • Bug 2695: Decision Tree - initial hiliting corrupted on single pattern stored for hiliting
  • Bug 2699: Windows 32bit build only comes with client VM
  • Bug 2704: Annotations produce huge DEBUG messages with empty lines
  • Bug 2710: XValidation statistics wrong
  • Bug 2718: Reference Row Filter does not allow filtering on RowID without columns
  • Bug 2719: File-Export Workflow throws odd exception if no workflow was selected
  • Bug 2720: External Tool framework has problems with empty input tables w/ chunking enabled
  • Bug 2722: PMML Predictor fails with NoSuchMethodException on MLP model
  • Bug 2725: Time To String node does not keep output column name (back to default after restart)
  • Bug 2738: Java Snippet Row Filter/Splitter do not assign ROWCOUNT field.
  • Bug 2740: Workspace not saved properly upon shutdown
  • Bug 2741: Postgresql databases ignore fetchsize when AUTOCOMMIT on
  • Bug 2742: Itemset Finder problem with non-integer values
  • Bug 2751: Table Creator node: "Input Line"" not editable
  • Bug 2757: XLS Reader should not save table spec
  • Bug 2680: Correlation View should use proper number formatting
  • Bug 2627: ColumnToGrid to group by certain selected column (better displaying cluster results)
  • Bug 2628: Static helper method for BooleanCell to return cell for boolean value
  • Bug 2687: Node description should allow tables
  • Bug 2712: DialogComponentFileChooser with flow variable buttons
  • Bug 2736: GroupBy aggregation method: percentage