Pdtable

Added

- [PR 95](https://github.com/startable/pdtable/pull/95) `write_excel()` with new parameter `style=True` applies hard-coded styles to table blocks in output workbook.

Added

- [92](https://github.com/startable/pdtable/issues/92) `write_excel()` can write to multiple sheets in a workbook.
- [85](https://github.com/startable/pdtable/issues/85) CSV and Excel writers can write transposed tables.

Changed

- [42](https://github.com/startable/pdtable/issues/42) Don't show dummy index when Table rendered as string.

Fixed

- Invalid attribute access to `TableBundle` (e.g. `bundle["nonexistent_table"]`) raised a dubious `KeyError`; now raises a more appropriate `AttributeError`.

Added

- Enable `in`-operator for TableBundle. Example: `assert "some_table_name" in bundle`

Fixed

- [87](https://github.com/startable/pdtable/issues/87) Parser fails on tables with zero rows. (Edge case not explicitly ruled out in the StarTable specification.)

Added

- For convenience, `Column` proxy is is now iterable. For example, `list(table["some_col"])` now works, whereas previously you had to access the backing `TableDataFrame` as `list(table.df["some_col"])`. Saving 3 chars FTW.
- CI pipeline now auto releases to PyPI on pushed tags starting with a "v".

Fixed

- \82: `read_excel()` fails on formulas

Added

- Major ergonomics improvement: The reader functions now output helpful error messages when parsing stops due to errors in the input table blocks.

Changed

- `write_excel()` parameter `path` renamed to `to`, reflecting the fact that it can write not only to file but also to binary stream; and by the same token, harmonizing its API with `read_csv()`.

Fixed

- 77: Harmful assert in `read_excel()` prevented reading from certain kinds of streams; now removed.