-
Notifications
You must be signed in to change notification settings - Fork 19
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Revert unproject_layout
's action; use shape_touched
in column optimization.
#317
Revert unproject_layout
's action; use shape_touched
in column optimization.
#317
Conversation
@@ -95,6 +95,7 @@ def test_json_bytes_no_delim_defined(ndjson_points_file: str) -> None: | |||
|
|||
|
|||
def test_to_and_from_dask_array(daa: dak.Array) -> None: | |||
daa = dak.from_awkward(daa.compute(), npartitions=3) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
the normal daa
fixture is an array that comes from parquet, and the form ends up getting a mask when we do column selection. We avoid the mask by going ahead and computing the parquet-originating collection and creating a new collection from the concrete result via from_awkward
. A bit roundabout but we still test what we mean to test here.
@@ -273,6 +274,8 @@ def test_from_lists(caa_p1: ak.Array) -> None: | |||
def test_to_dask_array(daa: dak.Array, caa: dak.Array) -> None: | |||
from dask.array.utils import assert_eq as da_assert_eq | |||
|
|||
daa = dak.from_awkward(daa.compute(), npartitions=4) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
same as above
Busy with other things today - but I can give it a shot later. |
of course always no rush :) |
1ab3212
to
b7f3b69
Compare
Codecov Report
❗ Your organization is not using the GitHub App Integration. As a result you may experience degraded service beginning May 15th. Please install the Github App Integration for your organization. Read more. @@ Coverage Diff @@
## main #317 +/- ##
==========================================
- Coverage 92.31% 90.82% -1.49%
==========================================
Files 21 21
Lines 2564 2562 -2
==========================================
- Hits 2367 2327 -40
- Misses 197 235 +38
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
b7f3b69
to
1482d37
Compare
@douglasdavis so this gets coffea tests to pass again but overtouching is has run rampant again in my usual torture tests, compared to dask-awkward 2023.7.0 + uproot 5.0.10 (where in my test with this branch I'm running this branch with uproot 5.0.10). |
Closing in favor of #319 |
Should fix #314 and #313.
This reverts the action that unproject_layout and plugs the hole left behind by using the
shape_touched
attribute of the typetracer report. The only negative side-effect we have is converting todask.array.Array
is failing in tests because of a mismatch between a regular numpy array and a masked numpy array.