petl-developers / petl / 38018819695

94%

push

github

Reuse the recast sort cache (#736)

`recast()` sorts the source again on each read of the same view. Keep
the sorted rows and discovered fields so later reads can reuse them.

Fixes #369.

## Changes

- Reuse the existing sort cache in memory or in temporary files.
- Preserve the sampled column order and lazy source reads. A failed sort
can be retried. Create a new view to read source changes after the cache
is filled.
- Add regression tests for repeat reads, source changes, partial reads,
iterator independence through `cat`, `stack`, and `annex`,
temporary-file cleanup, and source or reducer errors.
- Reject incompatible fields if another iterator fills the cache after a
header read.
- Document the cache behavior and add a change note.

Co-authored-by: Juarez Rudsatz <juarezr@gmail.com>

232 of 232 new or added lines in 2 files covered. (100.0%)

15210 of 16438 relevant lines covered (92.53%)

0.93 hits per line