Sling v1.5.14 (2026-04-06T15:41:03Z)

New Features

• SQL Server BCP with Entra ID authentication: BCP bulk loading now supports Entra ID token-based auth and Azure CLI integration for passwordless authentication.
• DuckDB Arrow IPC streaming: Implement Arrow IPC streaming for DuckDB connections, improving read/write performance.
• CDC full-refresh mode: Added support for full-refresh mode with CDC options, clearing state before running an initial snapshot.
• CDC UUID primary key support: CDC initial snapshots now handle string/UUID primary keys with offset-based chunking, in addition to integer PKs.
• length() function supports scalar types: The length() function now safely handles integers, floats, booleans, and dates by converting to string representation, allowing use in wildcard transforms with ternary guards.
• sling project jobs run --wait flag: New --wait / -w flag polls execution status every 5 seconds until the job reaches a terminal state, then reports rows, bytes, and duration.
• OAuth2 custom credentials and spec URL: Enhanced OAuth2 initialization to support custom client credentials and spec URL for flexible token management.

Bug Fixes

• Oracle sqlldr column sizing: Fixed sqlldr control file generation to use target column precision instead of hardcoding char(400000), preventing excessive memory allocation during Oracle bulk imports.
• TINYINT type handling: Corrected data type mapping for tinyint in metadata queries across MariaDB, MySQL, and StarRocks templates.
• Cross-platform path handling: Replaced path package with filepath across the codebase for correct behavior on Windows.
• Token save in agent mode: Allow local SaveToken to proceed even when running in agent mode.

Sling v1.5.13 (2026-03-24T13:18:25Z)

New Features

• OAuth2 authorization code flow with configurable redirect URL and port: The authorization code flow now supports a configured redirect_uri, using its host, port, and callback path instead of always binding to a random local port.
• Render OAuth2 scopes before configuration: OAuth2 scopes are now rendered (variable substitution) before being passed to the OAuth2 config, enabling dynamic scope values.
• CLIENT_SECRET support for Azure and OneLake connections: CLIENT_SECRET and client certificate properties are now properly passed through for Microsoft Fabric ABFS and Azure file system authentication.
• Unicode column name support: CSV header cleaning now preserves Unicode letters and numbers (e.g. Chinese, Japanese, accented characters) instead of stripping them.
• String length inference for file sources with column_typing: column_typing string max_length rules now apply to non-database sources (CSV/files), enabling auto-inference of varchar() precision when loading into databases like Oracle.

Bug Fixes

• BigQuery GetSchemata table retrieval: Fixed schema value not being cast to string, which caused table listing to fail when discovering BigQuery schemas.
• Oracle SQL*Loader out-of-memory handling: SQL*Loader now runs with context cancellation support (CommandContext) and surfaces actionable SQL*Loader-700 / out-of-memory errors instead of generic failures.
• Oracle SQL*Loader pipe error handling: Closed-pipe errors during CSV writing to SQL*Loader stdin are now caught and propagated correctly instead of being silently ignored.
• DuckDB connection leak during chunk loading: Added missing defer dbConn.Close() calls in ClearTableForChunkLoadWithRange and ProcessChunks to prevent connection leaks during chunked replication.
• Chunking error message without update_key: Fixed a spurious nil error being passed to g.Error when update_key is missing for stream chunking.
• Compact CDC State: Fixed CDC state so that GTID messages are stored compact

Sling v1.5.12 (2026-03-09T18:18:58Z)

New Features

• --home-dir flag: New flag across sling run and all sling conns subcommands (list, test, set, unset, exec, discover) to specify a custom sling home directory. Sets SLING_HOME_DIR and reloads connections accordingly.

• Column merge with defaults (+ prefix): Stream-level columns: entries prefixed with + now merge with defaults.columns instead of replacing them. Use +column_name: ~ to unset a default. Without +, legacy replace behavior is preserved.

• Pipeline store variables in replications: {store.xxx} variables set in pipeline steps are now accessible in inline replication where: and sql: fields.

• API spec range_step support: Added range_step to the endpoint context map, enabling three-part range expressions for backfill.

• API spec state variable rendering in jmespath / jq: State variables (e.g. {state.endpoint_name}) are now resolved inside jmespath and jq response record expressions.

Bug Fixes

• SLING_LOGGING env from replication: Moved SLING_LOGGING env handling from per-task to replication-level so it takes effect before stream execution and properly reinitializes the logger.

• Direct insert incremental state: writeToDbDirectly now calls SyncColumns() and SyncStats() so incremental state (e.g. datetime update keys) is correctly captured after direct insert writes.

• Auth token refresh on retry: API retry logic now refreshes HTTP request headers from conn.State.Auth.Headers before resending, fixing stale token issues after server-side token rotation (both single-request and mid-pagination scenarios).

• Windows path handling: CleanWindowsPath applied to local file locations in ParseLocation, API spec file:// paths, env file paths, and SLING_HOME_DIR.

• Oracle chunking with custom SQL: Fixed AS table alias syntax for Oracle (which doesn't support AS for subquery aliases).

Sling v1.5.11 (2026-03-02T19:52:22Z)

• Fix critical issue with API Spec state parsing when using setup

Sling v1.5.10 (2026-03-02T14:00:35Z)

Connectors

• ADBC DuckDB — new ADBC connector with Docker testing
• MongoDB — support ISODate() in filters
• ClickHouse — fix table rename SQL syntax
• SFTP — fix incremental file timestamp filtering
• Google Drive — add context timeout for API calls
• Local files — UNC path support (Windows network paths)

Connections & Environment

• Support setting username in connection URL
• New .env.sling file support for per-project connection definitions

MCP Server

• working_dir parameter on connection, pipeline, and replication actions
• Auto-loads .env.sling from working directory

CDC (Change Data Capture)

• Soft delete now uses _sling_synced_op = 'D' instead of _sling_deleted_at; snapshot rows marked with 'S'
• Updated merge_change_capture_soft templates across all 14 databases

Other

• Updated godbc v0.0.5, purego v0.10.0
• Disable AWS EC2 metadata service to prevent warning messages
• JMESPath/JQ safe rendering with state variables
• Simplified update check logic

Sling v1.5.9 (2026-02-23T15:13:21Z)

Change Data Capture (CDC)

• New change-capture replication mode
• Supports initial full/chunked snapshot then incremental CDC
• New merge strategies: change_capture and change_capture_soft (soft delete)
• CDC merge templates for: Postgres, MySQL, Snowflake, BigQuery, SQL Server, DuckDB, Redshift, Databricks, ClickHouse, SQLite, D1, DB2, Exasol, Oracle, StarRocks
• Pre-release for testing, not ready for production yet

Chunking Improvements

• Optional WHERE clause support added to chunk query functions
• Fix bug where start-date is 0000-01-01

CLI Cleanup

• Removed legacy subcommands
• Centralized execution ID into core/env package (env.ExecID)
• Enhanced file logging: exec ID prefix and log cleanup (added SLING_LOG_DIR env vars)

Proxy Support

• New SLING_PROXY environment variable (or proxy_url per-connection property) for proxy configuration

Bug Fixes

• StreamRowsContext column mapping streamlined; sync issue fixed
• Prometheus connector: consistent float conversion via toFloat method
• Time format strings in NewStreamProcessor clarified
• MySQL/MariaDB primary key handling improved in GenerateDDL
• InferColumnTypes preserves explicit types for empty datasets

Sling v1.5.8 (2026-02-17T11:17:57Z)

• Critical fix for AzureSQL and Fabric merge_strategy templates
• Improve S3 buffering to avoid back-pressure panic.

Sling v1.5.7 (2026-02-16T11:11:59Z)

Schema Migration

• Preserve foreign keys, indexes, column comments, default values, and auto-increment attributes when replicating between databases
• Supported for Oracle, PostgreSQL, Redshift, Snowflake, SQL Server, BigQuery, MySQL, DuckDB, ClickHouse, Databricks, SQLite
• Streams automatically ordered by foreign key dependencies
• Default value translation mappings between database engines

Merge Strategies

• Support for insert, update, update_insert, and delete_insert merge strategies

jq Expressions

• New jq source option for JSON-based data streams
• jq() transform function for column-level transformations
• Works with API sources and database JSON columns
• Compiled expressions cached for performance

Template & Expression Engine

• Nested template expression support for recursive resolution within nested braces (e.g., GraphQL queries)
• Safe rendering methods that keep missing expressions intact instead of erroring

Fixes & Improvements

• File-based logging setup for persistent log output
• Wildcard schema support in Discover method
• Duplicate step ID validation in pipeline configs
• DuckDB HTTP server shutdown fix after batch processing
• Total rows and bytes tracked in pipeline state
• Metadata queries updated to use current_database() for catalog filtering
• Zip file response handling for API sources

Sling v1.5.6 (2026-02-02T10:19:20Z)

• Enhanced delete_missing with configurable source and target WHERE clauses
• Added secret encoding support for HMAC authentication in API specs
• Added EscapeBackslash option for MySQL LOAD DATA operations
• Added row-start-only range format support for Excel

Bug Fixes

• Fixed Fabric ROWTERMINATOR to use hexadecimal representation
• Disabled AWS EC2 metadata retrieval for improved security
• Restored escape character handling in MySQL/MariaDB LOAD DATA INFILE templates

Sling v1.5.5 (2026-01-26T16:09:41Z)

New Features

• DB2 Database Support - Initial support for IBM DB2 with connection templates and full test suite
• definition-only Mode - Create table/file definitions without transferring data (parquet/arrow only for files)
• _sling_synced_at Column - Enable with SLING_SYNCED_AT_COLUMN=true for timestamp tracking
• _sling_synced_op Column - Tracks operation type: I (insert), U (update), D (soft delete)
• slugify Function - Convert strings to URL-friendly slugs

Bug Fixes

• DDL primary key with WITH clauses - Fixed incorrect placement when table options present (#694)
• Prometheus streaming deadlock - Fixed stall on large time ranges (#700)
• MySQL boolean CSV streaming - Fixed boolean handling in LOAD DATA LOCAL INFILE
• delete_missing with transforms - Fixed errors in PK-only queries when transforms present
• Connection caching - Fixed hash method and ODBC initialization

Improvements

• Go 1.25 - Upgraded build toolchain
• Docker ODBC packages - Added unixodbc and odbcinst
• Oracle XMLTYPE - Support for BigQuery transfers
• MySQL LOAD DATA LOCAL - Enhanced NULL value handling
• S3 multi-bucket access - Improved support in replications
• Schema discovery - Thread-safe schemata merging, column-level support