Result Batch V2 lets compatible planners return multiple result chunks in one response, reducing the number of fetch requests for large query results. This release candidate is published from vishal/plt-10376-result-batch-v2 for customer testing of PR #86 (PLT-10376).

Install the exact release candidate:

python -m pip install "e6data-python-connector==3.0.1rc1"

For native async use, install e6data-python-connector[async]==3.0.1rc1.

V2 remains disabled by default. Add enable_result_batch_v2=True to the existing connection configuration to test it against a planner that enables ENABLE_GET_NEXT_RESULT_BATCH_V2. Existing connection credentials and other settings are unchanged.

The connector preserves chunk order, delivers terminal-response rows, and falls back to V1 only when the server returns UNIMPLEMENTED. Receive limits remain unchanged, including the configurable 64 MiB async default. This release does not add background prefetching.

The feature branch passed the full offline Python 3.11-3.13 test matrix. A live SELECT 1 returned 1 over V2 with no V1 fallback and successful cleanup. Large-result performance, the 900-second workload requirement, and the broader live compatibility and failure matrix remain unqualified. This is a prerelease for testing, not a general-production rollout recommendation.

Source: #86

What's changed

• Native asyncio API for Python 3.11+: optional async connections, cursors, connection pooling, batch iteration and bounded cleanup through e6data_python_connector.aio.
• SQLAlchemy 2.0 async integration through the e6data+asyncio dialect and async-sqlalchemy extra.
• OAuth 2.0 authentication with client-credentials token acquisition, caching and renewal, optional scopes, Basic or POST client authentication, and externally supplied access tokens. OAuth sends a current bearer token on every RPC.
• Shared token lifecycle across matching sync and async connections, pools, threads and event loops within one process.
• Bounded cluster auto-resume after an exact suspended-cluster response during initial preparation, with authenticated status checks, at most one resume request and a bounded readiness wait. Query execution and result retrieval are not blindly replayed.
• Catalog discovery through synchronous Connection.get_catalogs().

Existing synchronous connections, pooling and username/password authentication remain available. Async support is optional.

Installation

# Base connector
python -m pip install --upgrade 'e6data-python-connector==3.0.0'

# Native asyncio API - Python 3.11+
python -m pip install --upgrade 'e6data-python-connector[async]==3.0.0'

# Async API with SQLAlchemy 2.0 - Python 3.11+
python -m pip install --upgrade 'e6data-python-connector[async-sqlalchemy]==3.0.0'

Usage and validation

Configure OAuth with secure=True and an HTTPS token endpoint. Supply exactly one authentication method. Externally supplied access tokens are not refreshed. OAuth and auto-resume require compatible, enabled server components.

The merged-commit CI passes all four offline test, coverage, wheel-build and fresh-install configurations. Live service qualification is a separate explicitly configured workflow and was not run as part of this publication.

The attached wheel and source archive were built from commit df85f81868aab047043b18a0eedf373fb2aae680. SHA256SUMS records their checksums.

See the Async API guide and OAuth lifecycle guide.

Implemented in PR #85.

Full changelog: 2.3.15...3.0.0

What's Changed

• fix(dialect): map engine short type names + strip type params in get_columns by @chandel-e6x in #84

New Contributors

• @chandel-e6x made their first contribution in #84

Full Changelog: 2.3.14...2.3.15

What's Changed

• README updated: Added status check for a query. by @vishale6x in #79
• fix dialect get_columns by @vishale6x in #83

Thanks to @arin-balyan1 for reporting the issue (#82) and providing the original fix (#81) that this release is based on.

Full Changelog: 2.3.13...2.3.14

What's Changed

• Windows support readme updated by @vishale6x in #77
• Fix/date year overflow jdbc parity by @vishale6x in #80

Full Changelog: 2.3.12...2.3.13

What's Changed

• Making fastbinary optional + SSL cert by @vishale6x in #78

Full Changelog: 2.3.11...2.3.12

What's Changed

• Decimal128Data, scale added. by @vishale6x in #73
• fix: Auto resume by @vishale6x in #74

Full Changelog: 2.3.10...2.3.11

What's Changed

• Proto File Updated by @AKASHSE-123 in #72
• Cluster name to lower case and Conn pool by @vishale6x in #71

Full Changelog: 2.3.9...2.3.10

What's Changed

• Switching strategy by @vishale6x in #68
• Query isolation. by @vishale6x in #70

Full Changelog: 2.3.8...2.3.9

What's Changed

• Doc updated. by @vishale6x in #67

Full Changelog: 2.3.7...2.3.8