Reference.
Keyboard shortcuts, supported formats, DataFusion SQL, variable interpolation, troubleshooting and limits of Odara ETL.
17. Appendix
17.1 Keyboard shortcuts
See the full reference in Section 3.6.
17.2 Supported formats
| Category | Read | Write |
|---|---|---|
| Files | CSV, Excel, XML, JSON, NDJSON, Parquet | CSV, Excel, XML, JSON, NDJSON, Parquet (Magic File auto-detects on read; writes .json as NDJSON) |
| Databases | PostgreSQL, MySQL, Oracle, SQL Server, MongoDB, ODBC, SQLite | Same as read |
| Data warehouses | Snowflake, ClickHouse, BigQuery, Redshift, DuckDB | Same as read |
| Cloud storage | S3, Azure Blob, Google Drive | Same as read |
| Remote servers | FTP, FTPS, SFTP | Same as read |
| APIs | REST (GET) | REST (POST/PUT/PATCH) |
17.3 DataFusion SQL reference
The SQL Transform uses Apache DataFusion, which implements most of ANSI SQL. Key features:
- Aggregates:
COUNT, SUM, AVG, MIN, MAX, ARRAY_AGG, APPROX_DISTINCT… - Window functions:
ROW_NUMBER, RANK, LAG, LEAD, FIRST_VALUE… with OVER / PARTITION BY / ORDER BY. - Joins: INNER, LEFT, RIGHT, FULL, SEMI, ANTI.
- Scalar functions: string (
concat, substr, regexp_match), date (date_trunc, date_part), numeric (abs, round, power), conditional (CASE, COALESCE, NULLIF). - CTEs:
WITH foo AS (…) SELECT … - Set ops: UNION, INTERSECT, EXCEPT.
17.4 Variable interpolation
Many fields support @variable interpolation, replaced at execution time with the current variable value (see Section 5.7). Supported in:
- File paths, directory paths.
- SQL queries (double-quoted to avoid conflict with column aliases).
- REST URLs, headers, query params, body templates.
- Email subject and body.
- Connection string fields.
Built-in variables: @date (today's date, ISO format), @datetime (current ISO timestamp), @run_id (unique ID of the current execution), @env (active environment name).
17.5 Troubleshooting
Pipeline execution returns 401 Unauthorized
Your session expired. Log in again; if using an API key from a script, verify the key is not revoked and the Authorization header uses the Bearer prefix.
Python Transform fails with "module not found"
The module isn't installed in Odara's Python environment. Add it to the node's Python Dependencies field — Odara runs pip install on first execution.
Oracle/DB2/ODBC nodes fail with driver errors
Odara uses native drivers that must be installed system-wide: Oracle Instant Client, IBM DB2 driver, or unixODBC / Windows ODBC. Use the Install Driver dialog (Help menu) to guide setup on common platforms.
Schedule doesn't fire
Check the timezone — cron is evaluated in the timezone bound to the schedule. Verify the pipeline/maestro still exists (deleted targets leave orphaned schedules).
17.6 Limits of this version
- Streaming: no streaming sources (Kafka, RabbitMQ, NATS, MQTT) in this build.
- Environment promotion: no Deployment Center (promoting pipelines from one environment to the next through the UI).
- Data observability: no lineage dashboard.
- Platforms: Linux x86_64 and Windows 64-bit; no macOS or ARM64 build yet.
- Talend Import Utility: in Beta (section 16).