Skip to content

CHANGELOG.md

Latest commit

 

History

History
202 lines (192 loc) · 13.3 KB

CHANGELOG.md

File metadata and controls

202 lines (192 loc) · 13.3 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Added

  • New feature: Support for macOS and Linux.
  • Documentation: Added API documentation in the Wiki.
  • New token_provider= parameter on connect() / Connection for Microsoft Entra ID authentication with a custom credential object. Accepts any object exposing a .get_token(scope) method (e.g. any azure-identity credential such as DefaultAzureCredential, AzureCliCredential, ManagedIdentityCredential). Mutually exclusive with Authentication= in the connection string and with attrs_before[SQL_COPT_SS_ACCESS_TOKEN]. Bulk copy re-acquires a fresh token from the provider on each operation. The token scope is fixed to the Azure commercial cloud; sovereign clouds are out of scope (supply a pre-acquired token via attrs_before instead).
  • GH-570: New Cursor.bulkcopy_arrow(table_name, source) method for high-performance bulk loading directly from Apache Arrow data. Accepts a pyarrow.Table, RecordBatch, or RecordBatchReader, any object exposing the Arrow C Data Interface (__arrow_c_stream__ / __arrow_c_array__ — e.g. polars, pandas 2.2+, DuckDB, ADBC results), or an iterable of record batches. Data is streamed to the server through the Arrow C Data Interface without materializing intermediate Python row objects, and the GIL is released for the duration of the network transfer. When the source data already originates as Arrow, this avoids the Arrow→tuple conversion the classic bulkcopy() path requires (measured ~1.4x–2.7x faster end-to-end for such sources). bulkcopy() now raises TypeError steering Arrow inputs to this method. Requires mssql-py-core 0.1.5+.
  • Bulk copy now supports Authentication=ActiveDirectoryServicePrincipal via an entra_id_token_factory callback registered on the mssql-py-core connection. The callback is invoked by mssql-tds mid-handshake (FedAuth workflow 0x02) so the tenant id can be resolved from the server-supplied STS URL. Requires mssql-py-core 0.1.5+. Partial fix for #534.
  • Standalone mssql-python-odbc package (PRs #663, #664): the Microsoft ODBC Driver 18 for SQL Server binaries are now also published as a separate, platform-specific mssql-python-odbc package. When it is installed, the native driver loader resolves the driver from it; when it is absent or incomplete, mssql-python transparently falls back to its own bundled libs/. This is a non-breaking step toward decoupling driver-binary updates from mssql-python releases; a future major version will make the dependency explicit and drop the bundled binaries.
  • Opt-in/opt-out native provider selection: a new mssql_python.native_provider module property (and MSSQL_PYTHON_NATIVE_PROVIDER environment variable, which takes precedence) lets a caller select which native ODBC provider is loaded: the default "msodbcsql18" (Microsoft ODBC Driver 18, unchanged behavior) or the opt-in "mssql-odbc" (a Rust-based driver, shipped inside the mssql-python-rs package alongside the Rust TDS core). The selection must be made before the first connection; it resolves and freezes then, and a later change is ignored with a RuntimeWarning; a conflicting property assignment also warns when the environment variable takes precedence. Call mssql_python.get_native_provider_info() for diagnostics (selected id, package, version, driver path, source, and whether it's frozen). This PR does not change the default provider or ship any Rust driver binaries.

Changed

  • Fetches reuse owned native metadata for stable columns within a result set; fetchmany() avoids the Python metadata-dictionary roundtrip. Re-execution, result transitions and statement/connection cleanup invalidate this metadata. Public descriptions stay fresh, name-validation timing is preserved, and sql_variant columns retain per-row descriptions and per-value probes. Fetch buffers, decoding settings and converted values are not cached.
  • DATE, TIME, and TIMESTAMP fetch conversion uses checked CPython constructors for the standard datetime types, while preserving cached substitute constructors, their positional arguments and exceptions, and fractional-second truncation. DATETIMEOFFSET, UUID, and Decimal conversion are unchanged.
  • mssql-python now depends on mssql-python-rs==0.1.0 for mssql_py_core instead of embedding files owned by that separately published distribution.
  • GH-769 deprecation policy: The misplaced GetInfoConstants members SQL_TXN_ISOLATION_LEVEL, SQL_CONCURRENCY, SQL_ROWSET_SIZE, SQL_ROW_NUMBER, SQL_IC_UPPER, SQL_IC_LOWER, SQL_IC_SENSITIVE, SQL_IC_MIXED, and SQL_SQL92_ENTRY_SQL, SQL_SQL92_INTERMEDIATE_SQL, SQL_SQL92_FULL_SQL remain available with their original values throughout 1.x. Existing enum attribute access, module-level imports, and get_info_constants() dictionary lookups remain available. Removal is deferred to 2.0 or later, only after maintainer approval and an explicit migration notice; this is not a scheduled removal. Deprecation is documented rather than emitting runtime warnings. These names are not information-type names, and passing their integers to getinfo() still requests unrelated information. They cannot be rejected by value without also rejecting legitimate information types with the same IDs. For migration, use SQL_ATTR_TXN_ISOLATION for the connection attribute; use ConstantsDDBC or module-level names for legacy statement options and SQL_IC_* response values, not as getinfo() requests. To interpret SQL_SQL_CONFORMANCE, use SQL_SC_SQL92_ENTRY (1), SQL_SC_FIPS127_2_TRANSITIONAL (2), SQL_SC_SQL92_INTERMEDIATE (4), and SQL_SC_SQL92_FULL (8). The deprecated SQL_SQL92_*_SQL names retain 127/128/129 solely for compatibility; they are not conformance flags. For information types whose IDs are corrected, previously persisted value-based enum pickles and raw IDs cannot identify their original meaning; rebuild them from the intended information-type names. Name-based enum pickles resolve retained names to their corrected values.
  • Connection strings and string connection parameters that contain a NUL (\x00) character are now rejected up front with InterfaceError instead of being silently truncated at the NUL by the underlying driver.
  • Improved error handling in the connection module.
  • GH-627 behavioral change: NULL parameters for VARBINARY/BINARY columns on physical tables now succeed silently (previously raised ProgrammingError when a non-NULL parameter was bound first). For temp tables where SQLDescribeParam cannot determine column metadata, the fallback to SQL_VARCHAR still produces the same ProgrammingError as before; users should call cursor.setinputsizes() to work around this.

Fixed

  • GH-754: Pooled connections are now rolled back and restored to autocommit mode before being parked. This prevents an empty transaction from remaining visible on an idle SQL Server session after Connection.close(). Abandoned native connections also roll back pending work before disconnecting during normal object destruction. Statement-handle allocation and cleanup are synchronized with disconnect, including cleanup invoked by cursor finalizers.
  • Bounded text fetched as UTF-16 now preserves leading U+FEFF and U+FFFE as payload rather than treating them as byte-order markers. This corrects row-wise fetchone(), fetchmany(), and fetchall() results, including bounded columns fetched alongside a MAX column, as well as bounded batch decoding on Linux/macOS. Existing platform-specific decode-error fallbacks remain unchanged. Actual MAX/LOB text decoding is unchanged; its existing BOM and trailing-NUL loss is not fixed here.
  • GH-745: executemany auto-detect binds money-range Decimal values as SQL_NUMERIC with a batch-wide precision/scale (still via SQL_C_CHAR string values), so a comparison against a smaller numeric column no longer overflows. SQL precision/scale stay on columnSize / decimalDigits; the CHAR array stride uses a separate bufferSize sized from the longest fixed-point encoding (e.g. Decimal("1E-38")), so near-max precision values are not rejected by the array buffer. The setinputsizes DECIMAL/NUMERIC string path uses the same buffer-width split; buffer width is derived from text produced by the protected conversion path so failed conversions still raise a sanitized ValueError with row/column details (no raw MemoryError/RuntimeError leakage).
  • GH-769: Corrected 11 GetInfoConstants IDs for scalar functions, outer joins, driver handles, cursor attributes, catalog support, and parameter descriptions. Added the ODBC name SQL_TIMEDATE_FUNCTIONS as an alias of SQL_DATETIME_FUNCTIONS at 52, preserving the existing canonical .name. For registered information types, getinfo() now uses ODBC return types instead of guessing from the bytes: numeric information is decoded as unsigned integers (including SQL_SQL_CONFORMANCE, SQL_CURSOR_SENSITIVITY, and SQL_MAX_IDENTIFIER_LEN), and character results use the Unicode path. An immutable return-type registry enforces the exact ODBC numeric width, including pointer-sized handles. Wrong-width or truncated numeric payloads raise DatabaseError rather than returning a guessed value. Legacy non-byte numeric payloads are not truncated or coerced from bool to int; convertible decimal strings still return integers. Also corrects decoding for the unlisted standard IDs SQL_DBMS_NAME (17), SQL_DBMS_VER (18), SQL_XOPEN_CLI_YEAR (10000), SQL_ASYNC_MODE (10021), and SQL_CREATE_ASSERTION (127). Other unlisted raw IDs retain their existing behavior. Driver Manager-only handle queries can still be unsupported by the native provider; correcting their IDs does not add Driver Manager support. Native retrieval errors, including timeout and connection-loss errors, continue to be logged and return None. Providers may return cached metadata after connection loss: getinfo() is not a connection-health check.
  • GH-740: A Python Decimal whose value falls in the SQL Server MONEY / SMALLMONEY range is now bound as SQL_NUMERIC with its own precision and scale on both execute() paths (native detection, and the legacy path reached when setinputsizes() covers fewer positions than parameters). Previously it was bound as a formatted VARCHAR based on the value alone, so comparing it against a smaller numeric/decimal column (WHERE v = ?) made SQL Server convert varchar→numeric and raise an arithmetic overflow instead of simply not matching. Also fixes a latent binder bug the shortcut was masking: the numeric APD descriptor record number was hardcoded to 1, so a numeric parameter in any position other than the first corrupted the parameter bound at position 1. Behavioral change on the wire: money-range Decimals now arrive as numeric rather than varchar — a bare SELECT ? returns Decimal instead of str, sql_variant stores them as numeric, and because numeric outranks money/varchar in data-type precedence, WHERE money_or_varchar_col = ? can add a CONVERT_IMPLICIT on the column side that turns an index seek into a scan. The executemany auto-detect path is covered separately by GH-745.
  • GH-725: The timeout parameter of connect() / Connection(...) now correctly sets the login (connection-attempt) timeout (SQL_ATTR_LOGIN_TIMEOUT), matching pyodbc and its own docstring. Previously it was silently applied as the per-statement query timeout, so connect(timeout=N) did not bound the connection attempt and instead aborted long-running queries. Behavioral change: the constructor timeout no longer affects Connection.timeout (the query timeout, still settable via the property, default 0); an explicit attrs_before[SQL_ATTR_LOGIN_TIMEOUT] takes precedence over the timeout kwarg. Timeout values are now validated consistently at both entry points: the constructor timeout and the Connection.timeout setter reject negative, non-integer, and bool values (previously Connection.timeout = True/False was accepted as 1/0).
  • GH-627: Fixed SQLDescribeParam ordinal remapping bug that caused VARBINARY/BINARY NULL bindings to fail when a non-NULL parameter was bound first. The driver now pre-resolves unknown NULL parameter types before any SQLBindParameter calls, avoiding ODBC ordinal confusion.
  • GH-726: The extension loader now derives the Windows architecture from sysconfig.get_platform() (the interpreter build) instead of platform.machine() (the host CPU). An x64 interpreter on a Windows ARM64 machine previously looked for ddbc_bindings.cpXY-arm64.pyd, missed, and took the fallback path on every import; it now resolves the amd64 binary the win_amd64 wheel ships. The fallback notice is now emitted as a RuntimeWarning instead of being printed to stdout.

[1.0.0-alpha] - 2025-02-24

Added

  • Initial release of the mssql-python driver for SQL Server.

Changed

  • N/A

Fixed

  • N/A