Arrow IPC binary fetch path for DataFrame execution#1489
Merged
Martozar merged 7 commits intogooddata:masterfrom Apr 13, 2026
Merged
Arrow IPC binary fetch path for DataFrame execution#1489Martozar merged 7 commits intogooddata:masterfrom
Martozar merged 7 commits intogooddata:masterfrom
Conversation
Martozar
requested review from
hkad98,
jaceksan,
lupko and
pcerny
as code owners
Martozar
force-pushed
the
c.mze-cq-105
branch
3 times, most recently
from
7453528 to
0380d40
Compare
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## master #1489 +/- ##
==========================================
+ Coverage 78.13% 78.66% +0.53%
==========================================
Files 228 230 +2
Lines 14926 15400 +474
==========================================
+ Hits 11662 12114 +452
- Misses 3264 3286 +22 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
hkad98
reviewed
Mar 30, 2026
no23reason
reviewed
Mar 31, 2026
Martozar
changed the title
no23reason
reviewed
Apr 1, 2026
no23reason
reviewed
Apr 1, 2026
Martozar
force-pushed
the
c.mze-cq-105
branch
2 times, most recently
from
d7fbc76 to
4e99271
Compare
Switch read_result_arrow to explicitly request application/vnd.apache.arrow.stream via Accept header and pipe the HTTP response directly into ipc.open_stream(), eliminating the intermediate BytesIO buffer. Update tests accordingly.
no23reason
previously approved these changes
Apr 1, 2026
Martozar
force-pushed
the
c.mze-cq-105
branch
3 times, most recently
from
2468c32 to
8dc2511
Compare
Add a parallel Arrow IPC execution path to DataFrameFactory and SeriesFactory that fetches results via the binary endpoint instead of JSON pagination: - arrow_convertor: pa.Table -> DataFrame conversion with label_overrides, grand_totals reordering, column_totals_indexes, primary_labels resolution, and metric field index helper - dataframe: for_exec_def_arrow(), for_arrow_table(), for_exec_result_id Arrow branch; Arrow path wired through for_visualization(), for_created_visualization() - series: use_arrow=True on indexed() / not_indexed() - ArrowConfig holds conversion params (self_destruct, types_mapper, custom_mapping); use_arrow is a dedicated DataFrameFactory.__init__ parameter risk: nonprod
Backfill column_totals_indexes into all 36 fixture meta.json files; extend parity tests to cover all four DataFrameMetadata fields (row_totals_indexes, column_totals_indexes, primary_labels_from_index, primary_labels_from_columns) and expand for_arrow_table tests from 4 hand-picked cases to the full fixture set. risk: nonprod
hkad98
reviewed
Apr 13, 2026
| # (C) 2026 GoodData Corporation | ||
| from __future__ import annotations | ||
|
|
||
| import json |
Contributor
There was a problem hiding this comment.
Consider using orjson – can be done as a follow-up.
hkad98
reviewed
Apr 13, 2026
| "python-dotenv~=1.0.0", | ||
| "deepdiff~=8.5.0", | ||
| "tests_support", | ||
| "pyarrow>=16.1.0", |
Contributor
There was a problem hiding this comment.
There is pyarrow>=23.0.1 in project.optional-dependencies. Consider unifying it.
Martozar
force-pushed
the
c.mze-cq-105
branch
2 times, most recently
from
40b16be to
3dc178e
Compare
Replace stdlib json with orjson in arrow_convertor.py for faster metadata parsing. Add orjson>=3.11.0 to the arrow optional dependency group and align the test group's pyarrow floor to match the arrow extra (>=23.0.1). risk: nonprod
hkad98
approved these changes
Apr 13, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds a native Arrow IPC binary fetch path to gooddata-pandas, providing a faster alternative to the existing JSON-paged AFM path for large result sets.
What changed
gooddata-sdk — binary fetch
gooddata-pandas — Arrow→DataFrame conversion
Why
The JSON paging path serialises every result to JSON and pages it in chunks — it is CPU-heavy and slow for wide or deep result sets. Arrow IPC transfers binary columnar
data in a single round-trip. End-to-end benchmarks against the GoodData demo workspace show 1.3×–33× speedup depending on table shape, with larger tables benefiting most .
Test coverage