Agate

-------------------------

This is a minor release fixing several small bugs that were blocking a downstream release of csvkit.

* Documented differing behavior of :class:`.MaxLength` under Python 2. (649)
* agate is now tested against Python 3.6. (650)
* Fix bug when :class:`.MaxLength` was called on an all-null column.
* Update extensions documentation to match new API. (645)
* Fix bug in :class:`.Change` and :class:`.PercentChange` where ``0`` values could cause ``None`` to be returned incorrectly.

-------------------------

This release adds SVG charting via the `leather <https://leather.rtfd.io>`_ charting library. Charts methods have been added for both :class:`.Table` and :class:`.TableSet`. (The latter create lattice plots.) See the revised tutorial and new cookbook entries for examples. Leather is still an early library. Please `report any bugs <https://github.com/wireservice/agate/issues>`_.

Also in this release are a :class:`.Slugify` computation and a variety of small fixes and improvements.

The complete list of changes is as follows:

* Remove support for monkey-patching of extensions. (594)
* :class:`.TableSet` methods which proxy :class:`.Table` methods now appear in the API docs. (640)
* :class:`.Any` and :class:`.All` aggregations no longer behave differently for boolean data. (636)
* :class:`.Any` and :class:`.All` aggregations now accept a single value as a test argument, in addition to a function.
* :class:`.Any` and :class:`.All` aggregations now require a test argument.
* Tables rendered by :meth:`.Table.print_table` are now GitHub Flavored Markdown (GFM) compatible. (626)
* The agate tutorial has been converted to a Jupyter Notebook.
* :class:`.Table` now supports ``len`` as a proxy for ``len(table.rows)``.
* Simple SVG charting is now integrated via `leather <https://leather.rtfd.io>`_.
* Added :class:`.First` computation. (634)
* :meth:`.Table.print_table` now has a `max_precision` argument to limit Number precision. (544)
* Slug computation now accepts an array of column names to merge. (617)
* Cookbook: standardize column values with :class:`.Slugify` computation. (613)
* Cookbook: slugify/standardize row and column names. (612)
* Fixed condition that prevents integer row names to allow bools in :meth:`.Table.__init__`. (627)
* :class:`.PercentChange` is now null-safe, returns None for null values. (623)
* :class:`.Table` can now be iterated, yielding :class:`Row` instances. (Previously it was necessarily to iterate :code:`table.rows`.)

--------------------

This release adds several new features, fixes numerous small bug-fixes, and improves performance for common use cases. There are some minor breaking changes, but few user are likely to encounter them. The most important changes in this release are:

1. There is now a :meth:`.TableSet.having` method, which behaves similarly to SQL's ``HAVING`` keyword.

2. :meth:`.Table.from_csv` is much faster. In particular, the type inference routines for parsing numbers have been optimized.

3. The :meth:`.Table.compute` method now accepts a ``replace`` keyword which allows new columns to replace existing columns "in place."" (As with all agate operations, a new table is still created.)

4. There is now a :class:`.Slug` computation which can be used to compute a column of slugs. The :meth:`.Table.rename` method has also added new options for slugifying column and row names.

The complete list of changes is as follows:

* Added a deprecation warning for ``patch`` methods. New extensions should not use it. (594)
* Added :class:`.Slug` computation (466)
* Added ``slug_columns`` and ``slug_rows`` arguments to :meth:`Table.rename`. (466)
* Added :meth:`.utils.slugify` to standardize a sequence of strings. (466)
* :meth:`.Table.__init__` now prints row and column on ``CastError``. (593)
* Fix null sorting in :meth:`.Table.order_by` when ordering by multiple columns. (607)
* Implemented configuration system.
* Fixed bug in :meth:`.Table.print_bars` when ``value_column`` contains ``None`` (608)
* :meth:`.Table.print_table` now restricts header on max_column_width. (605)
* Cookbook: filling gaps in a dataset with Table.homogenize. (538)
* Reduced memory usage and improved performance of :meth:`.Table.from_csv`.
* :meth:`.Table.from_csv` no longer accepts a sequence of row ids for :code:`skip_lines`.
* :meth:`.Number.cast` is now three times as fast.
* :class:`.Number` now accepts :code:`group_symbol`, :code:`decimal_symbol` and :code:`currency_symbols` arguments. (224)
* Tutorial: clean up state data under computing columns (570)
* :meth:`.Table.__init__` now explicitly checks that ``row_names`` are not ints. (322)
* Cookbook: CPI deflation, agate-lookup. (559)
* :meth:`.Table.bins` now includes values outside ``start`` or ``end`` in computed ``column_names``. (596)
* Fixed bug in :meth:`.Table.bins` where ``start`` or ``end`` arguments were ignored when specified alone. (599)
* :meth:`.Table.compute` now accepts a :code:`replace` argument that allows columns to be overwritten. (597)
* :meth:`.Table.from_fixed` now creates an agate table from a fixed-width file. (358)
* :mod:`.fixed` now implements a general-purpose fixed-width file reader. (358)
* :class:`TypeTester` now correctly parses negative currency values as Number. (595)
* Cookbook: removing a column (`select` and `exclude`). (592)
* Cookbook: overriding specific column types. (591)
* :class:`.TableSet` now has a :meth:`.TableSet._fork` method used internally for deriving new tables.
* Added an example of SQL's :code:`HAVING` to the cookbook.
* :meth:`.Table.aggregate` interface has been revised to be more similar to :meth:`.TableSet.aggregate`.
* :meth:`.TableSet.having` is now implemented. (587)
* There is now a better error when a forced column name does not exist. (591)
* Arguments to :meth:`.Table.print_html` now mirror :meth:`.Table.print_table`.

----------------------

The major feature of this release is new API documentation. Several minor features and bug fixes are also included. There are no major breaking changes in this release.

Internally, the agate codebase has been reorganized to be more modular, but this should be invisible to most users.

* The :class:`.MaxLength` aggregation now returns a `Decimal` object. (574)
* Fixed an edge case where datetimes were parsed as dates. (568)
* Fixed column alignment in tutorial tables. (572)
* :meth:`.Table.print_table` now defaults to printing ``20`` rows and ``6`` columns. (589)
* Added Eli Murray to AUTHORS.
* :meth:`.Table.__init__` now accepts a dict to specify partial column types. (580)
* :meth:`.Table.from_csv` now accepts a ``skip_lines`` argument. (581)
* Moved every :class:`.Aggregation` and :class:`.Computation` into their own modules. (565)
* :class:`.Column` and :class:`.Row` are now importable from `agate`.
* Completely reorgnized the API documentation.
* Moved unit tests into modules to match new code organization.
* Moved major :class:`.Table` and :class:`.TableSet` methods into their own modules.
* Fixed bug when using non-unicode encodings with :meth:`.Table.from_csv`. (560)
* :meth:`.Table.homogenize` now accepts an array of values as compare values if key is a single column name. (539)

-------------------------

This version implements several new features and includes two major breaking changes.

Please take note of the following breaking changes:

1. There is no longer a :code:`Length` aggregation. The more obvious :class:`.Count` is now used instead.

2. Agate's replacements for Python's CSV reader and writer have been moved to the :code:`agate.csv` namespace. To use as a drop-in replacement: :code:`from agate import csv`.

The major new features in this release are primarly related to transforming (reshaping) tables. They are:

1. :meth:`.Table.normalize` for converting columns to rows.
2. :meth:`.Table.denormalize` for converting rows to columns.
3. :meth:`.Table.pivot` for generating "crosstabs".
4. :meth:`.Table.homogenize` for filling gaps in data series.

Please see the following complete list of changes for a variety of other bug fixes and improvements.

* Moved CSV reader/writer to :code:`agate.csv` namespace.
* Added numerous new examples to the R section of the cookbook. (529-535)
* Updated Excel cookbook entry for pivot tables. (536)
* Updated Excel cookbook entry for VLOOKUP. (537)
* Fix number rendering in :meth:`.Table.print_table` on Windows. (528)
* Added cookbook examples of using :meth:`.Table.pivot` to count frequency/distribution.
* :meth:`.Table.bins` now has smarter output column names. (524)
* :meth:`.Table.bins` is now a wrapper around pivot. (522)
* :meth:`.Table.counts` has been removed. Use :meth:`.Table.pivot` instead. (508)
* :class:`.Count` can now count non-null values in a column.
* Removed :class:`.Length`. :class:`.Count` now works without any arguments. (520)
* :meth:`.Table.pivot` implemented. (495)
* :meth:`.Table.denormalize` implemented. (493)
* Added ``columns`` argument to :meth:`Table.join`. (479)
* Cookbook: Custom statistics/agate.Summary
* Added Kevin Schaul to AUTHORS.
* :meth:`Quantiles.locate` now correctly returns `Decimal` instances. (509)
* Cookbook: Filter for distinct values of a column (498)
* Added :meth:`.Column.values_distinct()` (498)
* Cookbook: Fuzzy phonetic search example. (207)
* Cookbook: Create a table from a remote file. (473)
* Added ``printable`` argument to :meth:`.Table.print_bars` to use only printable characters. (500)
* :class:`.MappedSequence` now throws an explicit error on __setitem__. (499)
* Added ``require_match`` argument to :meth:`.Table.join`. (480)
* Cookbook: Rename columns in a table. (469)
* :meth:`.Table.normalize` implemented. (487)
* Added :class:`.Percent` computation with example in Cookbook. (490)
* Added Ben Welsh to AUTHORS.
* :meth:`.Table.__init__` now throws a warning if auto-generated columns are used. (483)
* :meth:`.Table.__init__` no longer fails on duplicate columns. Instead it renames them and throws a warning. (484)
* :meth:`.Table.merge` now takes a ``column_names`` argument to specify columns included in new table. (481)
* :meth:`.Table.select` now accepts a single column name as a key.
* :meth:`.Table.exclude` now accepts a single column name as a key.
* Added :meth:`.Table.homogenize` to find gaps in a table and fill them with default rows. (407)
* :meth:`.Table.distinct` now accepts sequences of column names as a key.
* :meth:`.Table.join` now accepts sequences of column names as either a left or right key. (475)
* :meth:`.Table.order_by` now accepts a sequence of column names as a key.
* :meth:`.Table.distinct` now accepts a sequence of column names as a key.
* :meth:`.Table.join` now accepts a sequence of column names as either a left or right key. (475)
* Cookbook: Create a table from a DBF file. (472)
* Cookbook: Create a table from an Excel spreadsheet.
* Added explicit error if a filename is passed to the :class:`.Table` constructor. (438)

------------------------

This release adds several minor features. The only breaking change is that default column names will now be lowercase instead of uppercase. If you depended on these names in your scripts you will need to update them accordingly.

* :class:`.TypeTester` no longer takes a ``locale`` argument. Use ``types`` instead.
* :class:`.TypeTester` now takes a ``types`` argument that is a list of possible types to test. (461)
* Null conversion can now be disabled for :class:`.Text` by passing ``cast_nulls=False``. (460)
* Default column names are now lowercase letters instead of uppercase. (464)
* :meth:`.Table.merge` can now merge tables with different columns or columns in a different order. (465)
* :meth:`.MappedSequence.get` will no longer raise ``KeyError`` if a default is not provided. (467)
* :class:`.Number` can now test/cast the ``long`` type on Python 2.