ibis.ibis_to_pq

ibis.ibis_to_pq(
    table,
    out_file,
    *,
    engine=None,
    row_group_size=1024 * 1024,
    threads=None,
    tz='UTC',
    adbc_batch_size_hint_bytes=None,
    adbc_use_copy=None,
    **writer_kwargs,
)

Write an Ibis PostgreSQL table expression to a parquet file.

This helper compiles an Ibis PostgreSQL expression to SQL and runs it through the same PostgreSQL export engines used elsewhere in db2pq. The resulting Arrow stream is written directly to the destination Parquet file.

ibis_to_pq() currently supports Ibis expressions backed by a PostgreSQL connection. To use it, install the optional dependency:

pip install "db2pq[ibis]"

Parameters

Name	Type	Description	Default
table		Ibis table expression backed by PostgreSQL. This may be a base table or a derived expression such as a filtered, selected, or mutated query.	required
out_file	str or path - like	Destination parquet file path.	required
engine	(duckdb, adbc)	Query execution engine used to run the compiled PostgreSQL SQL. If omitted, uses the configured default engine from set_default_engine() / DB2PQ_ENGINE.	"duckdb"
row_group_size	int	Maximum number of rows in each written Parquet row group.	1024 * 1024
threads	int	Maximum DuckDB worker threads to use when engine="duckdb".	None
tz	str	Time zone assumption for naive PostgreSQL timestamps before normalizing Parquet output to UTC.	'UTC'
adbc_batch_size_hint_bytes	int	ADBC batch size hint in bytes when engine="adbc".	None
adbc_use_copy	bool	Explicitly enable or disable the PostgreSQL ADBC driver’s COPY optimization when engine="adbc".	None
**writer_kwargs		Additional keyword arguments passed to pyarrow.parquet.ParquetWriter. This can be used to set options such as compression="zstd".	{}

Returns

Name	Type	Description
pq_file	str	Name of parquet file created.

Raises

Name	Type	Description
	TypeError	If the supplied Ibis expression is not backed by PostgreSQL, or if PostgreSQL connection information cannot be determined from the backend.

Examples

>>> from db2pq import ibis_to_pq
>>> expr = con.table("my_table").filter(lambda t: t.id > 100)
>>> ibis_to_pq(expr, "my_table.parquet")
'my_table.parquet'

>>> expr = con.table("my_table").select("id", "value")
>>> ibis_to_pq(expr, "my_table.parquet", compression="zstd")
'my_table.parquet'