feat(duckdb): Amazon S3 / Object Storage data source (DTL-1808)#2785
Draft
m1n0 wants to merge 2 commits into
Draft
feat(duckdb): Amazon S3 / Object Storage data source (DTL-1808)#2785m1n0 wants to merge 2 commits into
m1n0 wants to merge 2 commits into
Conversation
Add an `object_storage` block to the DuckDB connection config so a `type: duckdb` source can point at Amazon S3. Its presence marks the source as S3-backed: on connect the engine installs/loads httpfs, resolves AWS credentials (static access-key or STS AssumeRole, with optional ExternalId), registers a DuckDB S3 secret (CREATE SECRET, with a SET s3_* pragma fallback for DuckDB < 0.10), and exposes each configured dataset (prefix+glob) as a lazy view over read_parquet / read_csv_auto / read_json_auto. Views appear in information_schema.tables so the existing discovery/profiling/scan pipeline reuses them unchanged. Also fix the file-extension connection branch, which silently ignored the `configuration` dict: it now applies the settings via SET (the `:default:` connection cannot accept a config= kwarg). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Object-storage (S3) datasets are exposed as DuckDB VIEWs, but the default discovery/metadata path (discover_qualified_objects with object_types=None) enumerates TABLEs only, so real profiling/discovery found nothing for an S3 source. DuckDBDataSourceImpl.discover_qualified_objects now includes VIEWs in the default object types when the source is an object-storage source. Standard and local-file DuckDB sources (which materialize TABLEs) keep the TABLE-only default, so their behavior is unchanged. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Amazon S3 / Object Storage data source (DTL-1808)
Adds S3 / object-storage support to the DuckDB engine plugin so files in S3 (Parquet/CSV/JSON) can be scanned directly, before ingestion into a warehouse.
What's in it
connection.object_storageblock on the DuckDB data source:provider: s3,region, polymorphicauth(assume-role or access-key), anddatasets(each a name + S3 path/glob + format).INSTALL/LOAD httpfs, build S3 credentials (STS AssumeRole via boto3, or static access-key),CREATE SECRET, and oneCREATE VIEWper dataset overread_parquet/read_csv_auto/read_json_auto(globs supported).configuration-dict-ignored gap.Tests
41 passing in
soda-duckdb— config parse (both auth modes), view creation over local files, assume-role (STS mocked), discovery, error cases.Runtime note
The runner image must have the DuckDB
httpfsextension available (egress at connect time, or pre-bundled offline).Part of the cross-repo object-storage feature (
soda,soda-contract-launcher,docs). Ticket: DTL-1808.