0.5.0
Introduced the Bus
, a Series
-like container of mulitple Frame
, supporting lazily reading from and writing to XLSX, SQLite, and HDF5 data stores, as well as zipped pickles and delimited files.
Added interface
attribute to all containers, providing a hierarchical presentation of all interfaces.
Added display_tall()
and display_wide()
convenience methods to all containers.
Added label_widths_at_depth()
on Index
and IndexHierarchy
.
Added Series.from_concat_items()
and Frame.from_concat_items()
.
Added Frame.to_xarray()
.
Added Frame.to_xlsx()
, Frame.from_xlsx()
.
Added Frame.to_sqlite()
, Frame.from_sqlite()
.
Added Frame.to_hdf5()
, Frame.from_hdf5()
.
Added Frame.to_rst()
.
Added Frame.to_markdown()
.
Added Frame.to_latex()
.
The interface of Frame.from_delimited
(as well as Frame.from_csv
and Frame.from_tsv
) has been updated to conform to the common usage of index_depth
and columns_depth
. IndexHierarchy is now supported when index_depth
or columns_depth
is greater than one. The former parameter index_column
is renamed index_column_first
.
Added IndexHierarchy.from_index_items
and IndexHierarchy.from_labels_delimited
.
Added IndexBase.names
attribute to provide normalized names equal in length to depth.
The DisplayConfig
parameter type_show
now, if False, hides, native class types used as headers. This is the default display for all specialized string output via Frame.to_html
, Frame.to_rst
, Frame.to_markdown
, Frame.to_latex
, as well as Jupyter display methods.
Added Frame.unset_index()
.
Added Frame.pivot()
.
Added Frame.iter_window
, Frame.iter_window_items
, Frame.iter_window_array
, Frame.iter_window_array_items
.
Added Series.iter_window
, Series.iter_window_items
, Series.iter_window_array
, Series.iter_window_array_items
.
Added Frame.bloc
and Frmae.assign.bloc
Added IndexHierarchy.rehierarch
, Series.rehierarch
, and Frame.rehierarch
.
Defined __bool__
for all containers, where the result is determined based on if the underlying NumPy array has size
greater than zero.
Improved Frame.to_pandas()
to preserve columnar types.
Frame.set_index_hierarchy
now accepts a reorder_for_hierarchy
argument, reordering the rows to support hierarchability.
Added Frame.from_dict_records
and Frame.from_dict_records_items
; when given records, the union of all keys is used to derive columns.