Data Inlining is currently experimental. It needs to be enabled explicitly and is only supported for DuckDB databases. We are planning to improve support for this feature in the future.

When writing small changes to DuckLake, it can be wasteful to write each changeset to an individual Parquet file. DuckLake supports directly writing small changes to the metadata using Data Inlining. Instead of writing a Parquet file to the data storage and then writing a reference to that file in the metadata catalog, we directly write the rows to inlined data tables within the metadata catalog.

Data inlining must be enabled explicitly using the DATA_INLINING_ROW_LIMIT attach parameter. When enabled, any inserts that write fewer than the given amount of rows are automatically written to inlined tables instead.

ATTACH 'ducklake:inlining.db' (DATA_INLINING_ROW_LIMIT 10);

Inlined data behaves exactly the same as data written to Parquet files. The inlined data can be queried, updated and deleted, and the schema of inlined data can be modified.
The only difference is that the inlined data lives in the metadata catalog, instead of in Parquet files in the data path.

For example, when inserting a low number of rows, data is automatically inlined:

CREATE TABLE inlining.tbl(col INTEGER);
-- inserting 3 rows, data is inlined
INSERT INTO inlining.tbl VALUES (1), (2), (3);
-- no Parquet files exist
SELECT COUNT(*) FROM glob('inlining.db.files/**');

┌──────────────┐
│ count_star() │
│    int64     │
├──────────────┤
│      0       │
└──────────────┘

When inserting more data than the DATA_INLINING_ROW_LIMIT, inserts are automatically written to Parquet:

INSERT INTO inlining.tbl FROM range(100);
SELECT COUNT(*) FROM glob('inlining.db.files/**');

┌──────────────┐
│ count_star() │
│    int64     │
├──────────────┤
│      1       │
└──────────────┘

Flushing Inlined Data

There is currently no built-in functionality for flushing inlined data to Parquet files. We plan to implement this soon.