Version 0.25.0

Made Column and ColumnType take parameter arguments reflecting the type of data held in each specialized column/type.

v0.24.9

• Fixed fillMissing methods and added corresponding unit test

v0.24.8

• DoubleColumn methods now return DoubleColumn instead of NumberColumn

v0.24.7

• Added full outer joins

v0.24.6

• Added methods for filling missing values
• Improved Row performance

Version 0.24.5

Remove's the plot sub-project.

People who need visualization with Tablesaw are encouraged to use the tablesaw-jsplot library instead. This layer is already better in many ways than the original. See the visualization section of the Tablesaw user guide for more information.

The native java plotting code was always something of a mess given the lack of a single, strong, appropriately licensed open-sourcevisualization library in Java. It relied on several other libraries to piece together its functionality, including Smile, XCharts, and JavaFx.

The JavaFx dependency was particularly concerning since it is large, and was moved to a separate module in Java 9, and removed from the JDK entirely in Java 11, presenting some additional complexity from a build/deployment standpoint across JDK versions.

Beyond that, the functionality of the library was very limited and both the API and the visual user interface were inconsistent due to the underlying inconsistencies between the libraries it was built on.

Version 0.24.4

This release contains several enhancements, and considerable refactoring, some of which is not backward compatible. It also contains Codacy and travis updates and documentation improvements.

Refactoring:

• replaced ColumnType.NUMBER with ColumnType.DOUBLE. It was a bad decision to name the type NUMBER, and this walk-balk of that mistake was inevitable. If this affects your code, a search/replace of NUMBER with DOUBLE should fix.
• refactored a large amount of column-type related functionality into a revised ColumnType interface, including moving from enum to class-based column types and removing many places where code switched on column type. The work required to add a new column type is now much clearer and more centralized, although additional work is needed.
• the above refactoring included a rationalization of column-specific CSV parsing, which had become quite a mess, as it was spread out in many places and included a bunch of special case logic.

Enhancements:

• Extended Joins to allow the client to specify whether duplicate column names should be allowed. If allowed, the duplicate columns are renamed; if not, an exception is thrown if any columns in table 2 has the same name as a column in table 1.
• made column lookup by name case-insensitive in Row. Now it works like column-lookup in Tables.

Docs:

• Extension of Time Series example code

Version 0.24.3

Enhancements, documentation, refactoring, and bug fixes in Visualization

Version 0.24.1

v0.24.1

[maven-release-plugin] copy for tag v0.24.1

Version 0.24.0

Bug fixes:

• fix for dateTimeFormatter issue #313 and extended formatting fix to time and date columns.
• fix for bug introduced in fix for #229. CrossTab counts were not rendering totals
• if the first category type was numeric.

Enhancements:

• Simplified scatter and histogram classes and extended plot types for canned plots

• Improved file handling on Plots, each plot gets its own temp file

• Introduce a number of new canned plot types

• Extensive improvements to visualization documentation.

• Provided a converter to Smile's AttributeDataset (#311)

• restored helpful method from removed interface IntConvertibleColumn

• restored useful DoubleArray utils not in TableConverter