Choose chDB Python Arrow input, result, insert, or standard Arrow database-driver paths when a task uses PyArrow tables or readers. Not for ordinary SQL or DataStore work.
82%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
Passed
No findings from the security scan
Arrow Database Connectivity (ADBC) is a separate standardized driver API. Python(tbl) is chDB's table function for a Python object already in the calling application; a RecordBatchReader is a one-shot source of batches.
pyarrow.Table: chdb.query("SELECT ... FROM Python(tbl)"). For an incoming one-shot pyarrow.RecordBatchReader, use Python(reader); do not promise a second scan. Neither requires constructing a separate session or using ADBC.chdb.query("INSERT INTO dst SELECT ... FROM Python(tbl)", path=data_path); keep the source alive through the query. For several operations on one in-memory database, keep a session rather than opening one per call.chdb.query_arrow(sql) returns a pyarrow.Table; with chdb.query_arrow_batches(sql) as batches: yields incremental RecordBatch values and closes a partial result.chdb[adbc] is the appropriate separate driver route. Its BindStream on a general SELECT binds parameters, not a temporary table in FROM; prefer Python(tbl) / Python(reader) for a query over Python objects. Start with docs/guides/querying-apache-arrow.mdx for the Python Arrow documentation map and examples.chdb-io/chdb
f85dbd3
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.