rich7420 commented on PR #1393:
URL: https://github.com/apache/mahout/pull/1393#issuecomment-4678507935
@0lai0 one follow-up before we merge 🙏
Heads up: a **null outer list row** makes the whole read fail. Reading a
`List<Float64>` of `[[1,2], null, [3,4]]` returns:
```
InvalidInput("Inconsistent sample sizes: expected 2, got 0")
```
`FillZero` doesn't help — a null row's `value_length(i)` is 0, which trips
the consistency check (and if it's row 0, `sample_size` locks to 0).
- It's **pre-existing** (old per-row code did the same), **not introduced by
this PR**
- But real-world nullable List columns hit it, and there's no null-outer-row
test
- Related: `written / self.sample_size.unwrap_or(1)` — `unwrap_or(1)` only
guards `None`, not `Some(0)`. Unreachable in practice (the check errors first),
but a cheap `Some(0)` guard wouldn't hurt
Not blocking — would you be up for a follow-up (null-row test + clearer
error, or just skip null rows)? Or fine to track it as a known limitation under
#1338 👍
Everything else passes locally — 332 tests green, clippy/fmt clean, round-1
comments all addressed. LGTM overall.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
