petl-developers / petl / 39295358161

94%

push

github

Fix toarray() on tables without data rows (#746)

This fixes `toarray()` / `torecarray()` on a
table that has a header but no data rows, e.g. `etl.toarray([('foo',
'bar')])`. With no `dtype`, the type is inferred with `np.rec.array()`
over the sampled rows, which raises `IndexError: tuple index out of
range` when there are none.

Changes:

1. Fixed a bug in `infer_dtype()`: with no rows it now returns a dtype
with the header's field names and `object` type (which is also what
pandas gives for an empty DataFrame), so an empty array is returned.
2. Added a test and a changelog entry.

35 of 35 new or added lines in 2 files covered. (100.0%)

15533 of 16759 relevant lines covered (92.68%)

0.93 hits per line