Skip to content

Q IPC publisher for kdb Insights Enterprise databases.

Use DBPublisher to publish data directly to a kdb Insights Enterprise database over a Q IPC connection to the TP port.

Supports CSV, Parquet, JSON files, directories, Pandas DataFrames, and pykx Table objects.

Note

Requires pykx. Install with: pip install "kxi[all]"

Example 
import pandas as pd
from kxi.publish.dbpublisher import DBPublisher

df = pd.DataFrame({"sym": ["AAPL", "GOOG"], "price": [150.0, 175.0]})

with DBPublisher(host="localhost", port=5010) as pub:
    pub.publish(df, table="quotes")
    pub.publish("/path/to/trades.csv", table="trades")

Classes:

  • DBPublisher – Q IPC publisher for a kdb Insights Enterprise database.

DBPublisher

DBPublisher(host, port, user=None, password=None, chunksize=10000, pub_fn='.u.pub')

Q IPC publisher for a kdb Insights Enterprise database.

Connects to a kdb+ database's TP (tickerplant) port over Q IPC and inserts data into named tables. Use as a context manager to ensure the connection is closed after publishing.

Example 
import pandas as pd
from kxi.publish.dbpublisher import DBPublisher

df = pd.DataFrame({"sym": ["AAPL", "GOOG"], "price": [150.0, 175.0]})

with DBPublisher(host="localhost", port=5010) as pub:
    pub.publish(df, table="quotes")
    pub.publish("/path/to/trades.csv", table="trades")

Functions:

  • close – Close the Q IPC connection to the kdb Insights database.
  • publish – Publish data to a named table on the kdb Insights database.

Connect to a kdb Insights database TP port over Q IPC.

Parameters:

  • host (str) – Hostname or IP address of the kdb Insights TP server.
  • port (int) – Port number for the Q IPC connection.
  • user (Optional[str]) – Username (not used in current version).
  • password (Optional[str]) – Password (not used in current version).
  • chunksize (int) – Number of rows to read per chunk when publishing files. Avoids loading entire files into memory. Default: 10000. Not supported for columnar JSON (file_format="json").
  • pub_fn (str) – Q function to call for publishing. Default: ".u.pub".

Raises:

close

close()

Close the Q IPC connection to the kdb Insights database.

publish

publish(data, table, *, file_format=None, type_map=None, **kwargs)

Publish data to a named table on the kdb Insights database.

Parameters:

  • data (str | Path | DataFrame) – Data source. Accepted types:

  • File path string or pathlib.Path (e.g. "/data/trades.csv")

  • HTTP, S3, or GCS URL (e.g. "s3://bucket/data.parquet")
  • Directory path — all direct child files are published
  • pandas.DataFrame
  • pykx.Table
  • table (str) – Destination table name in the kdb Insights database.
  • file_format (DataFormat | str | None) – File format override: csv, parquet, json, or json_records. Auto-detected from file extension if not set.
  • type_map (dict | None) – Override pandas type casting for specific columns. Maps column names to target types: "timedelta", "datetime", or "numeric". Example: {"time": "timedelta", "ts": "datetime"}
  • *kwargs* – Additional arguments forwarded to the Q publish function.
Back to top