petl-developers / petl / 43492268764

94%

push

github

Handle missing and mixed-type pivot categories (#811)

`pivot()` raises `TypeError` before yielding a header when category
values cannot be directly compared, for example an ordinary missing
category alongside a string:

```python
import petl as etl

source = [('region', 'category', 'sales'),
          ('north', None, 3), ('north', 'retail', 10),
          ('south', 'retail', 7)]
list(etl.pivot(source, 'region', 'category', 'sales', sum))
# TypeError: '<' not supported between instances of 'str' and 'NoneType'
```

The source sort already accepts these values. The failure is the
separate native sort of the discovered output categories.

## Changes

Use the existing `Comparable` key to order pivot categories, consistent
with the mixed-type ordering used by `sort()` and `recast()`. The
example then produces:

```python
[('region', None, 'retail'), ('north', 3, 10), ('south', None, 7)]
```

48 of 48 new or added lines in 2 files covered. (100.0%)

16832 of 17950 relevant lines covered (93.77%)

0.94 hits per line