This updates the flux integration to use the `flux/array` package rather
than directly using the `arrow/array` package.
Flux recently switched to wrapping the array types from arrow and
creating its own array package to be used for table columns instead of
directly referencing the arrow package. This allows us to keep a
consistent interface, but potentially change internal implementations
without changing downstream consumers. Most recently, the
`*array.String` type has some of its own optimizations for certain array
patterns.
This change updates the flux integration to use the new API.
* fix: field metaqueries take fast path if predicate is only on `_measurement`
* chore: update CHANGELOG
* test: add test for fields with measurement predicate
The kind for experimental-to is in conflict with the one used by native
flux. In the past, native flux didn't use the kind and expected
implementations to register their own procedure spec with that value.
In the latest Flux, this isn't the case and it is using the kind for a
native implementation. As we have not implemented the requirements for
the native implementation, we are still replacing the procedure spec but
using a different name that's unique to our implementation.
* feat(query): enable min/max pushdown
* fix(query): fix the group last pushdown to use descending cursors
* test(storage): add read group test with no agg
Co-authored-by: Jonathan A. Sternberg <jonathan@influxdata.com>
* fix(storage): Detect need for descending cursor in WindowAggregate
* chore: Format + comments
* chore: PR cosmetic feedback (#21141)
Co-authored-by: Phil Bracikowski <pbracikowski+git@influxdata.com>
* chore: rename testcase and fix comments
Co-authored-by: Phil Bracikowski <pbracikowski+git@influxdata.com>
A static initialization is not desirable in the main binaries, as it forces all
paths of code to init, but it is still useful in tests. It allows static
intialization to be performed once for all tests and eliminates the need to
always add the FluxInit call. Added a fluxinit/static package that calls
fluxinit.FluxInit() to replace the builtin package. This hides the nature of
the initialization and makes it clear that it is mandatory initialization code
getting called.
It appears that the double write caused by using to() inside a separate
execution environment (experimental.chain) causes flux e2e tests to behave
unpredictably, when coupled with the 1.x storage engine. Removing the second
write by using two passes, one to write to the db, then another to run the
test, eliminates the flakiness. Verified by running e2e tests in parallel times
8 for 12 hours without any flakiness observed. Before the fix, the flakiness
would take approx 30 minutes on avgerage to exhibit.
This commit also removes universe/to_time from the skipped tests because it was
added when this flakiness was discovered.
This is a backport of #14262 to the 1.x storage engine. The 1.x storage
engine is now the primary engine for open source so when we switched we
regressed to the old behavior.
This also fixes `go generate` for the tsm1 package by running `tmpl`
with `go run` instead of assuming the correct one is installed in the
path.
This is required to keep the system resources low when running
the Flux end-to-end tests, which create a bucket for each test. A
bucket creates at least 17 files after the first write:
* 8 for the `_series` segment files
* 8 for the `index` log files
* 1 for the `wal`
Enables the mix and max aggregates for the ReadGroupAggregte pushdown behind a feature flag.
Co-authored-by: Jonathan A. Sternberg <jonathan@influxdata.com>
Daniel Moran
Jonathan A. Sternberg
William Baker
Christopher M. Wolff
Sam Arnold
Faith Chikwekwe
Paul Hummer
Sean Brickley
Paul Hummer
Adrian Thurston
Yiqun (Ethan) Zhang
jl
Faith Chikwekwe
Stuart Carnie
Brett Buddin
Paul Hummer
George
jlapacik