dlt-altertable 0.1.0__tar.gz

dlt_altertable-0.1.0/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

dlt_altertable-0.1.0/PKG-INFO

@@ -0,0 +1,234 @@
+Metadata-Version: 2.4
+Name: dlt-altertable
+Version: 0.1.0
+Summary: dlt destination that loads into the Altertable lakehouse through its HTTP API
+Keywords: dlt,altertable,lakehouse,parquet
+Author: Altertable
+Author-email: Altertable <hello@altertable.ai>
+License-Expression: Apache-2.0
+License-File: LICENSE
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: Topic :: Database
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Requires-Dist: dlt[parquet]~=1.19
+Requires-Dist: pyarrow>=15.0.0
+Requires-Dist: requests>=2.26
+Requires-Dist: urllib3>=2
+Requires-Python: >=3.12
+Project-URL: Homepage, https://altertable.ai
+Project-URL: Repository, https://github.com/altertable-ai/dlt-altertable
+Project-URL: Changelog, https://github.com/altertable-ai/dlt-altertable/blob/main/CHANGELOG.md
+Description-Content-Type: text/markdown
+
+# dlt-altertable
+
+[![CI](https://github.com/altertable-ai/dlt-altertable/actions/workflows/ci.yml/badge.svg)](https://github.com/altertable-ai/dlt-altertable/actions/workflows/ci.yml)
+[![PyPI](https://img.shields.io/pypi/v/dlt-altertable)](https://pypi.org/project/dlt-altertable/)
+[![Python 3.12+](https://img.shields.io/badge/python-3.12+-3776AB.svg)](https://www.python.org)
+[![License: Apache 2.0](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](LICENSE)
+
+A [dlt](https://dlthub.com) destination that loads into the [Altertable](https://altertable.ai)
+lakehouse through the HTTP Lakehouse API. Each dlt load job is one parquet file, posted verbatim
+to `/upload` or `/upsert`. An append or replace file written before dlt evolved the load's schema
+is padded with typed `NULL` columns first, because those uploads must match the table column for
+column. Merge files always post as they are, so columns a file omits keep their stored values.
+
+## Install
+
+```bash
+uv add dlt-altertable
+```
+
+Supported: Python 3.12 to 3.14, `dlt >= 1.19`. The only other dependencies are `pyarrow`, `requests`, and `urllib3`.
+
+## Use
+
+```python
+import dlt
+from dlt_altertable import altertable
+
+
+@dlt.resource(
+    name="contacts",
+    write_disposition="merge",
+    primary_key="id",
+    columns={"lastmodifieddate": {"dedup_sort": "desc"}},
+)
+def contacts():
+    yield [
+        {"id": 1, "email": "ada@example.com", "lastmodifieddate": 100},
+        {"id": 2, "email": "grace@example.com", "lastmodifieddate": 100},
+    ]
+
+
+pipeline = dlt.pipeline(pipeline_name="crm", destination=altertable)
+pipeline.run(contacts())
+```
+
+Configure the connection in `.dlt/secrets.toml`:
+
+```toml
+[destination.altertable]
+host = "api.altertable.ai"
+catalog = "lakehouse"
+dataset_name = "crm"
+username = "..."
+password = "..."
+# port = 443, tls = true and compute_size = "XS" (for schema queries) are the defaults
+```
+
+Every setting can also come from dlt's environment variables, which override the toml file
+(`DESTINATION__ALTERTABLE__HOST`, `DESTINATION__ALTERTABLE__PASSWORD`, and so on), or be passed
+directly: `altertable(host=..., catalog=..., dataset_name=...)`.
+
+No configuration is needed in an environment that already exports the `ALTERTABLE_HOST`,
+`ALTERTABLE_PORT`, `ALTERTABLE_TLS`, `ALTERTABLE_USERNAME`, `ALTERTABLE_PASSWORD`,
+`ALTERTABLE_CATALOG` and `ALTERTABLE_SCHEMA` variables. The host must be the HTTP API host
+(`api.…`), not the Flight SQL endpoint.
+
+Two notes on naming:
+
+- The pipeline's `dataset_name` is not used. dlt does not route it to custom destinations, so the
+  target schema is the destination's own `dataset_name` setting.
+- To point several pipelines at different backends, name each configuration:
+  `altertable(destination_name="altertable_staging")` reads `[destination.altertable_staging]`.
+
+## Write dispositions
+
+| dlt write disposition | HTTP call                                | Notes                                                |
+| --------------------- | ---------------------------------------- | ---------------------------------------------------- |
+| `append`              | `POST /upload?mode=append`               | The destination creates or evolves the table first.  |
+| `replace`             | `mode=overwrite`, then `mode=append`     | The first file of each load recreates the table.     |
+| `merge`               | `POST /upsert?primary_key=…&cursor_field=…` | Server-side upsert on the `primary_key`.          |
+
+Merge always runs as a server-side upsert on the `primary_key`, dlt's `upsert` merge strategy.
+A resource that does not name a strategy is accepted and upserted the same way. Configurations
+the upsert cannot honor fail the load with a terminal error instead of loading under different
+semantics: an explicit `delete-insert` or `scd2` strategy, a `merge_key`, a `hard_delete`
+column, a missing `primary_key`, or an ascending `dedup_sort`.
+
+A merge that sends a subset of columns updates only those columns on matched rows; omitted
+columns keep their current values in the lakehouse.
+
+### Picking the winning row
+
+When two rows collide on the primary key, Altertable keeps the one with the highest value of the
+column marked with dlt's standard [`dedup_sort` hint](https://dlthub.com/docs/general-usage/merge-loading):
+
+```python
+@dlt.resource(
+    write_disposition="merge",
+    primary_key="id",
+    columns={"lastmodifieddate": {"dedup_sort": "desc"}},
+)
+```
+
+Altertable arbitrates against existing table rows as well as within the batch, a superset of
+dlt's batch-only deduplication. Without the hint, which row wins is up to the server.
+
+## dlt system columns and tables
+
+Like dlt's SQL destinations, this destination loads dlt's lineage columns: every row carries
+`_dlt_id` and `_dlt_load_id`, making it traceable to the load that produced it. The dataset also
+gains a `_dlt_pipeline_state` table. dlt maintains `_dlt_loads` and `_dlt_version` only inside
+SQL destinations, so those tables do not appear here.
+
+## Schema evolution
+
+The destination owns the target table's schema. Before each load it syncs the table through
+`POST /query`: a missing table is created with typed DDL derived from dlt's schema, and when
+dlt's schema evolution adds a column, the destination issues `ALTER TABLE ... ADD COLUMN`, so
+existing tables follow the source. Removed columns stay in the table and keep their values.
+Columns that only ever contained `NULL` are dropped by dlt at normalize time with a warning,
+before they reach the destination.
+
+Nested data does not become child tables: the destination sets `max_table_nesting=0`, dlt's
+default for custom destinations, so lists and objects land as JSON strings in the parent table.
+
+## Atomicity and retries
+
+Each parquet file is one HTTP POST, applied atomically by the server: a file either lands fully
+or not at all. A load split across several files is not atomic as a whole, and the first file of
+a `replace` load recreates the table.
+
+Failures map onto dlt's retry contract: errors a retry cannot fix, such as bad credentials or
+an invalid request, fail the job immediately with a terminal error, while server errors,
+capacity timeouts and connection failures are retried by dlt until the fifth attempt fails
+(`load.raise_on_max_retries`). Errors streamed mid-response by `/query` retry as well, which
+is safe because every statement the destination issues is `IF NOT EXISTS`-idempotent. The upload
+read timeout is one hour, because the server answers only once the file is fully ingested. Do
+not lower it: a request aborted client side can still complete on the server, which turns dlt's
+retry into duplicate appended rows. `merge` tables are idempotent under retry, `append` is at
+least once, and `replace` is at least once after its first file, since only the first file of
+the load recreates the table and a retried later file appends twice.
+
+The destination declares `loader_parallelism_strategy="table-sequential"`, and the `replace`
+bookkeeping depends on it. Do not override it with `LOAD__PARALLELISM_STRATEGY=parallel`: two
+files of one replace load would both recreate the table and silently lose rows.
+
+One caveat discovered the hard way: the target schema is created on demand, so a typo in
+`dataset_name` does not fail, it lands data in a new schema. The catalog, by contrast,
+must exist.
+
+## Incremental state on ephemeral runners
+
+[Custom destinations cannot restore pipeline state](https://dlthub.com/docs/dlt-ecosystem/destinations/destination),
+so incremental cursors live only in dlt's own state under `pipelines_dir`. On an ephemeral runner
+that directory is gone on the next run, and every load starts from scratch.
+
+Either persist `pipelines_dir` between runs, or bootstrap the cursor from the destination itself
+before extracting:
+
+```python
+import json
+import requests
+
+response = requests.post(
+    "https://api.altertable.ai/query",
+    json={"statement": f'SELECT max(lastmodifieddate) FROM "{catalog}"."{schema}"."{table_name}"'},
+    auth=(username, password),
+    timeout=(10, 300),
+)
+response.raise_for_status()
+rows = [json.loads(line) for line in response.text.splitlines()]
+if any(isinstance(row, dict) and "error" in row for row in rows):
+    raise RuntimeError(rows)
+since = rows[2][0]
+```
+
+The response streams one JSON line per row after two metadata lines, so the first row sits at
+index 2. A failed query arrives as an `{"error": ...}` line inside the stream, not as a special
+row, so check for it before trusting the value. An empty table gives a `NULL` max, which simply
+means start from the beginning.
+
+## Develop
+
+```bash
+uv run --locked pytest
+uv run --locked ruff check .
+uv run --locked ruff format --check .
+uvx ty check src
+```
+
+The test suite mocks the HTTP API and needs no server. For a live check, run
+[altertable-mock](https://github.com/altertable-ai/altertable-mock) and point a pipeline at its
+Lakehouse REST port:
+
+```bash
+docker run -d -p 15100:15000 \
+  -e ALTERTABLE_MOCK_LAKEHOUSE_PORT=15000 -e ALTERTABLE_MOCK_USERS="dlt-demo:lk_demo" \
+  ghcr.io/altertable-ai/altertable-mock:latest
+```
+
+## Resources
+
+- [dlt custom destination docs](https://dlthub.com/docs/dlt-ecosystem/destinations/destination)
+- [dlt merge loading](https://dlthub.com/docs/general-usage/merge-loading)
+- [Altertable](https://altertable.ai)
+
+## License
+
+Apache 2.0, see [LICENSE](LICENSE).

dlt_altertable-0.1.0/README.md

@@ -0,0 +1,209 @@
+# dlt-altertable
+
+[![CI](https://github.com/altertable-ai/dlt-altertable/actions/workflows/ci.yml/badge.svg)](https://github.com/altertable-ai/dlt-altertable/actions/workflows/ci.yml)
+[![PyPI](https://img.shields.io/pypi/v/dlt-altertable)](https://pypi.org/project/dlt-altertable/)
+[![Python 3.12+](https://img.shields.io/badge/python-3.12+-3776AB.svg)](https://www.python.org)
+[![License: Apache 2.0](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](LICENSE)
+
+A [dlt](https://dlthub.com) destination that loads into the [Altertable](https://altertable.ai)
+lakehouse through the HTTP Lakehouse API. Each dlt load job is one parquet file, posted verbatim
+to `/upload` or `/upsert`. An append or replace file written before dlt evolved the load's schema
+is padded with typed `NULL` columns first, because those uploads must match the table column for
+column. Merge files always post as they are, so columns a file omits keep their stored values.
+
+## Install
+
+```bash
+uv add dlt-altertable
+```
+
+Supported: Python 3.12 to 3.14, `dlt >= 1.19`. The only other dependencies are `pyarrow`, `requests`, and `urllib3`.
+
+## Use
+
+```python
+import dlt
+from dlt_altertable import altertable
+
+
+@dlt.resource(
+    name="contacts",
+    write_disposition="merge",
+    primary_key="id",
+    columns={"lastmodifieddate": {"dedup_sort": "desc"}},
+)
+def contacts():
+    yield [
+        {"id": 1, "email": "ada@example.com", "lastmodifieddate": 100},
+        {"id": 2, "email": "grace@example.com", "lastmodifieddate": 100},
+    ]
+
+
+pipeline = dlt.pipeline(pipeline_name="crm", destination=altertable)
+pipeline.run(contacts())
+```
+
+Configure the connection in `.dlt/secrets.toml`:
+
+```toml
+[destination.altertable]
+host = "api.altertable.ai"
+catalog = "lakehouse"
+dataset_name = "crm"
+username = "..."
+password = "..."
+# port = 443, tls = true and compute_size = "XS" (for schema queries) are the defaults
+```
+
+Every setting can also come from dlt's environment variables, which override the toml file
+(`DESTINATION__ALTERTABLE__HOST`, `DESTINATION__ALTERTABLE__PASSWORD`, and so on), or be passed
+directly: `altertable(host=..., catalog=..., dataset_name=...)`.
+
+No configuration is needed in an environment that already exports the `ALTERTABLE_HOST`,
+`ALTERTABLE_PORT`, `ALTERTABLE_TLS`, `ALTERTABLE_USERNAME`, `ALTERTABLE_PASSWORD`,
+`ALTERTABLE_CATALOG` and `ALTERTABLE_SCHEMA` variables. The host must be the HTTP API host
+(`api.…`), not the Flight SQL endpoint.
+
+Two notes on naming:
+
+- The pipeline's `dataset_name` is not used. dlt does not route it to custom destinations, so the
+  target schema is the destination's own `dataset_name` setting.
+- To point several pipelines at different backends, name each configuration:
+  `altertable(destination_name="altertable_staging")` reads `[destination.altertable_staging]`.
+
+## Write dispositions
+
+| dlt write disposition | HTTP call                                | Notes                                                |
+| --------------------- | ---------------------------------------- | ---------------------------------------------------- |
+| `append`              | `POST /upload?mode=append`               | The destination creates or evolves the table first.  |
+| `replace`             | `mode=overwrite`, then `mode=append`     | The first file of each load recreates the table.     |
+| `merge`               | `POST /upsert?primary_key=…&cursor_field=…` | Server-side upsert on the `primary_key`.          |
+
+Merge always runs as a server-side upsert on the `primary_key`, dlt's `upsert` merge strategy.
+A resource that does not name a strategy is accepted and upserted the same way. Configurations
+the upsert cannot honor fail the load with a terminal error instead of loading under different
+semantics: an explicit `delete-insert` or `scd2` strategy, a `merge_key`, a `hard_delete`
+column, a missing `primary_key`, or an ascending `dedup_sort`.
+
+A merge that sends a subset of columns updates only those columns on matched rows; omitted
+columns keep their current values in the lakehouse.
+
+### Picking the winning row
+
+When two rows collide on the primary key, Altertable keeps the one with the highest value of the
+column marked with dlt's standard [`dedup_sort` hint](https://dlthub.com/docs/general-usage/merge-loading):
+
+```python
+@dlt.resource(
+    write_disposition="merge",
+    primary_key="id",
+    columns={"lastmodifieddate": {"dedup_sort": "desc"}},
+)
+```
+
+Altertable arbitrates against existing table rows as well as within the batch, a superset of
+dlt's batch-only deduplication. Without the hint, which row wins is up to the server.
+
+## dlt system columns and tables
+
+Like dlt's SQL destinations, this destination loads dlt's lineage columns: every row carries
+`_dlt_id` and `_dlt_load_id`, making it traceable to the load that produced it. The dataset also
+gains a `_dlt_pipeline_state` table. dlt maintains `_dlt_loads` and `_dlt_version` only inside
+SQL destinations, so those tables do not appear here.
+
+## Schema evolution
+
+The destination owns the target table's schema. Before each load it syncs the table through
+`POST /query`: a missing table is created with typed DDL derived from dlt's schema, and when
+dlt's schema evolution adds a column, the destination issues `ALTER TABLE ... ADD COLUMN`, so
+existing tables follow the source. Removed columns stay in the table and keep their values.
+Columns that only ever contained `NULL` are dropped by dlt at normalize time with a warning,
+before they reach the destination.
+
+Nested data does not become child tables: the destination sets `max_table_nesting=0`, dlt's
+default for custom destinations, so lists and objects land as JSON strings in the parent table.
+
+## Atomicity and retries
+
+Each parquet file is one HTTP POST, applied atomically by the server: a file either lands fully
+or not at all. A load split across several files is not atomic as a whole, and the first file of
+a `replace` load recreates the table.
+
+Failures map onto dlt's retry contract: errors a retry cannot fix, such as bad credentials or
+an invalid request, fail the job immediately with a terminal error, while server errors,
+capacity timeouts and connection failures are retried by dlt until the fifth attempt fails
+(`load.raise_on_max_retries`). Errors streamed mid-response by `/query` retry as well, which
+is safe because every statement the destination issues is `IF NOT EXISTS`-idempotent. The upload
+read timeout is one hour, because the server answers only once the file is fully ingested. Do
+not lower it: a request aborted client side can still complete on the server, which turns dlt's
+retry into duplicate appended rows. `merge` tables are idempotent under retry, `append` is at
+least once, and `replace` is at least once after its first file, since only the first file of
+the load recreates the table and a retried later file appends twice.
+
+The destination declares `loader_parallelism_strategy="table-sequential"`, and the `replace`
+bookkeeping depends on it. Do not override it with `LOAD__PARALLELISM_STRATEGY=parallel`: two
+files of one replace load would both recreate the table and silently lose rows.
+
+One caveat discovered the hard way: the target schema is created on demand, so a typo in
+`dataset_name` does not fail, it lands data in a new schema. The catalog, by contrast,
+must exist.
+
+## Incremental state on ephemeral runners
+
+[Custom destinations cannot restore pipeline state](https://dlthub.com/docs/dlt-ecosystem/destinations/destination),
+so incremental cursors live only in dlt's own state under `pipelines_dir`. On an ephemeral runner
+that directory is gone on the next run, and every load starts from scratch.
+
+Either persist `pipelines_dir` between runs, or bootstrap the cursor from the destination itself
+before extracting:
+
+```python
+import json
+import requests
+
+response = requests.post(
+    "https://api.altertable.ai/query",
+    json={"statement": f'SELECT max(lastmodifieddate) FROM "{catalog}"."{schema}"."{table_name}"'},
+    auth=(username, password),
+    timeout=(10, 300),
+)
+response.raise_for_status()
+rows = [json.loads(line) for line in response.text.splitlines()]
+if any(isinstance(row, dict) and "error" in row for row in rows):
+    raise RuntimeError(rows)
+since = rows[2][0]
+```
+
+The response streams one JSON line per row after two metadata lines, so the first row sits at
+index 2. A failed query arrives as an `{"error": ...}` line inside the stream, not as a special
+row, so check for it before trusting the value. An empty table gives a `NULL` max, which simply
+means start from the beginning.
+
+## Develop
+
+```bash
+uv run --locked pytest
+uv run --locked ruff check .
+uv run --locked ruff format --check .
+uvx ty check src
+```
+
+The test suite mocks the HTTP API and needs no server. For a live check, run
+[altertable-mock](https://github.com/altertable-ai/altertable-mock) and point a pipeline at its
+Lakehouse REST port:
+
+```bash
+docker run -d -p 15100:15000 \
+  -e ALTERTABLE_MOCK_LAKEHOUSE_PORT=15000 -e ALTERTABLE_MOCK_USERS="dlt-demo:lk_demo" \
+  ghcr.io/altertable-ai/altertable-mock:latest
+```
+
+## Resources
+
+- [dlt custom destination docs](https://dlthub.com/docs/dlt-ecosystem/destinations/destination)
+- [dlt merge loading](https://dlthub.com/docs/general-usage/merge-loading)
+- [Altertable](https://altertable.ai)
+
+## License
+
+Apache 2.0, see [LICENSE](LICENSE).

dlt_altertable-0.1.0/pyproject.toml

@@ -0,0 +1,68 @@
+[project]
+name = "dlt-altertable"
+version = "0.1.0"
+description = "dlt destination that loads into the Altertable lakehouse through its HTTP API"
+readme = "README.md"
+license = "Apache-2.0"
+license-files = ["LICENSE"]
+keywords = [
+    "dlt",
+    "altertable",
+    "lakehouse",
+    "parquet",
+]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Intended Audience :: Developers",
+    "Topic :: Database",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Programming Language :: Python :: 3.14",
+]
+requires-python = ">=3.12"
+dependencies = [
+    "dlt[parquet]~=1.19",
+    "pyarrow>=15.0.0",
+    "requests>=2.26",
+    "urllib3>=2",
+]
+
+[[project.authors]]
+name = "Altertable"
+email = "hello@altertable.ai"
+
+[project.urls]
+Homepage = "https://altertable.ai"
+Repository = "https://github.com/altertable-ai/dlt-altertable"
+Changelog = "https://github.com/altertable-ai/dlt-altertable/blob/main/CHANGELOG.md"
+
+[dependency-groups]
+dev = [
+    "pytest~=8.4",
+    "ruff~=0.14",
+]
+
+[build-system]
+requires = ["uv_build>=0.11.29,<0.12.0"]
+build-backend = "uv_build"
+
+[tool.ruff]
+line-length = 100
+target-version = "py312"
+
+[tool.ruff.lint]
+select = [
+    "E",
+    "F",
+    "I",
+    "UP",
+    "B",
+]
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]
+addopts = [
+    "-v",
+    "--tb=short",
+    "--strict-markers",
+]

dlt_altertable-0.1.0/pyproject.toml.orig

@@ -0,0 +1,48 @@
+[project]
+name = "dlt-altertable"
+version = "0.1.0"
+description = "dlt destination that loads into the Altertable lakehouse through its HTTP API"
+readme = "README.md"
+authors = [{ name = "Altertable", email = "hello@altertable.ai" }]
+license = "Apache-2.0"
+license-files = ["LICENSE"]
+keywords = ["dlt", "altertable", "lakehouse", "parquet"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Intended Audience :: Developers",
+    "Topic :: Database",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Programming Language :: Python :: 3.14",
+]
+requires-python = ">=3.12"
+
+dependencies = [
+    "dlt[parquet]~=1.19",
+    "pyarrow>=15.0.0",
+    "requests>=2.26",
+    "urllib3>=2",
+]
+
+[project.urls]
+Homepage = "https://altertable.ai"
+Repository = "https://github.com/altertable-ai/dlt-altertable"
+Changelog = "https://github.com/altertable-ai/dlt-altertable/blob/main/CHANGELOG.md"
+
+[dependency-groups]
+dev = ["pytest~=8.4", "ruff~=0.14"]
+
+[build-system]
+requires = ["uv_build>=0.11.29,<0.12.0"]
+build-backend = "uv_build"
+
+[tool.ruff]
+line-length = 100
+target-version = "py312"
+
+[tool.ruff.lint]
+select = ["E", "F", "I", "UP", "B"]
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]
+addopts = ["-v", "--tb=short", "--strict-markers"]

dlt_altertable-0.1.0/src/dlt_altertable/__init__.py

@@ -0,0 +1,3 @@
+from dlt_altertable.destination import altertable
+
+__all__ = ["altertable"]

dlt_altertable-0.1.0/src/dlt_altertable/api.py

@@ -0,0 +1,80 @@
+import json
+from http import HTTPStatus
+from typing import Any
+
+import requests
+from dlt.common.destination.exceptions import DestinationTerminalException
+from requests.adapters import HTTPAdapter
+
+from dlt_altertable.configuration import AltertableClientConfiguration
+
+UPLOAD_TIMEOUT = (30, 3600)
+QUERY_TIMEOUT = (10, 300)
+UPLOAD_WRITE_BLOCK_BYTES = 1 << 20
+LOADER_WORKERS = 20
+TERMINAL_STATUSES = {
+    HTTPStatus.BAD_REQUEST,
+    HTTPStatus.UNAUTHORIZED,
+    HTTPStatus.PAYMENT_REQUIRED,
+    HTTPStatus.FORBIDDEN,
+    HTTPStatus.NOT_FOUND,
+    HTTPStatus.METHOD_NOT_ALLOWED,
+}
+
+
+class LargeBlockAdapter(HTTPAdapter):
+    """A load job posts a whole parquet file in one stream, and urllib3's default blocksize is
+    sized for ordinary requests, not for a file-sized body."""
+
+    def init_poolmanager(self, *args: Any, **kwargs: Any) -> None:
+        kwargs["blocksize"] = UPLOAD_WRITE_BLOCK_BYTES
+        super().init_poolmanager(*args, **kwargs)
+
+
+session = requests.Session()
+_adapter = LargeBlockAdapter(pool_maxsize=LOADER_WORKERS)
+session.mount("http://", _adapter)
+session.mount("https://", _adapter)
+
+
+def raise_for_failure(response: requests.Response, action: str) -> None:
+    if response.status_code == HTTPStatus.OK:
+        return
+    detail = f"{action} failed with HTTP {response.status_code}: {response.text.strip()[:2000]}"
+    if response.status_code in TERMINAL_STATUSES:
+        raise DestinationTerminalException(detail)
+    raise RuntimeError(detail)
+
+
+def execute_sql(config: AltertableClientConfiguration, statement: str) -> list[list]:
+    response = session.post(
+        f"{config.base_url}/query",
+        json={"statement": statement, "ephemeral": True, "compute_size": config.compute_size},
+        auth=config.basic_auth,
+        timeout=QUERY_TIMEOUT,
+    )
+    raise_for_failure(response, f"query {statement!r}")
+    payload = [json.loads(line) for line in response.text.splitlines() if line.strip()]
+    for entry in payload:
+        if isinstance(entry, dict) and "error" in entry:
+            raise RuntimeError(f"query {statement!r} failed mid-stream: {entry['error']}")
+    return payload[2:]
+
+
+def post_parquet(
+    config: AltertableClientConfiguration,
+    endpoint: str,
+    params: dict[str, str],
+    parquet_file_path: str,
+    action: str,
+) -> None:
+    with open(parquet_file_path, "rb") as parquet_file:
+        response = session.post(
+            f"{config.base_url}/{endpoint}",
+            params=params,
+            data=parquet_file,
+            auth=config.basic_auth,
+            headers={"Content-Type": "application/parquet"},
+            timeout=UPLOAD_TIMEOUT,
+        )
+    raise_for_failure(response, action)

dlt_altertable-0.1.0/src/dlt_altertable/configuration.py

@@ -0,0 +1,54 @@
+import os
+from typing import cast
+
+from dlt.common.configuration import ConfigurationValueError, configspec
+from dlt.common.typing import TSecretStrValue
+from dlt.destinations.impl.destination.configuration import CustomDestinationClientConfiguration
+
+ENV_FALLBACKS = {
+    "host": "ALTERTABLE_HOST",
+    "catalog": "ALTERTABLE_CATALOG",
+    "dataset_name": "ALTERTABLE_SCHEMA",
+    "username": "ALTERTABLE_USERNAME",
+    "password": "ALTERTABLE_PASSWORD",
+}
+
+
+@configspec
+class AltertableClientConfiguration(CustomDestinationClientConfiguration):
+    """Falls back to the `ALTERTABLE_*` environment variables after dlt's own providers,
+    so a preconfigured environment needs no explicit destination configuration."""
+
+    host: str | None = None
+    catalog: str | None = None
+    dataset_name: str | None = None
+    username: str | None = None
+    password: TSecretStrValue | None = None
+    port: int | None = None
+    tls: bool | None = None
+    compute_size: str | None = None
+
+    def on_resolved(self) -> None:
+        for parameter, env_var in ENV_FALLBACKS.items():
+            value = getattr(self, parameter) or os.environ.get(env_var)
+            if not value:
+                raise ConfigurationValueError(
+                    f"{parameter} is not configured: pass {parameter}= to altertable(), set "
+                    f"destination.altertable.{parameter} in .dlt/secrets.toml, or export "
+                    f"{env_var}."
+                )
+            setattr(self, parameter, value)
+        if self.port is None:
+            self.port = int(os.environ.get("ALTERTABLE_PORT", "443"))
+        if self.tls is None:
+            self.tls = os.environ.get("ALTERTABLE_TLS", "true").lower() != "false"
+        if self.compute_size is None:
+            self.compute_size = os.environ.get("ALTERTABLE_COMPUTE_SIZE", "XS")
+
+    @property
+    def base_url(self) -> str:
+        return f"{'https' if self.tls else 'http'}://{self.host}:{self.port}"
+
+    @property
+    def basic_auth(self) -> tuple[str, str]:
+        return cast(tuple[str, str], (self.username, self.password))

dlt_altertable-0.1.0/src/dlt_altertable/destination.py

@@ -0,0 +1,114 @@
+from typing import cast
+
+import dlt
+from dlt.common.destination.exceptions import DestinationTerminalException
+from dlt.common.schema import TTableSchema
+from dlt.common.schema.utils import (
+    get_columns_names_with_prop,
+    get_dedup_sort_tuple,
+    has_column_with_prop,
+)
+
+from dlt_altertable.api import post_parquet
+from dlt_altertable.configuration import AltertableClientConfiguration
+from dlt_altertable.table_schema import aligned_parquet, create_or_evolve_table
+
+
+def primary_key_columns(table: TTableSchema) -> list[str]:
+    return get_columns_names_with_prop(table, "primary_key")
+
+
+def unsupported_merge_configuration(table: TTableSchema) -> str | None:
+    strategy = table.get("x-merge-strategy")
+    if strategy not in (None, "upsert"):
+        return f"merge strategy {strategy!r}"
+    if has_column_with_prop(table, "merge_key"):
+        return "merge_key"
+    if has_column_with_prop(table, "hard_delete"):
+        return "the hard_delete column hint"
+    dedup_sort = get_dedup_sort_tuple(table)
+    if dedup_sort and dedup_sort[1] != "desc":
+        return f"dedup_sort {dedup_sort[1]!r} (the server keeps the highest value, use 'desc')"
+    if not primary_key_columns(table):
+        return "merge without a primary_key"
+    return None
+
+
+def upsert_params(table: TTableSchema) -> dict[str, str] | None:
+    if table.get("write_disposition") != "merge":
+        return None
+    if unsupported := unsupported_merge_configuration(table):
+        raise DestinationTerminalException(
+            f"Table {table['name']}: {unsupported} is not supported by the Altertable "
+            "destination, which runs merge as a server-side upsert on the primary_key."
+        )
+    dedup_sort = get_dedup_sort_tuple(table)
+    cursor = dedup_sort[0] if dedup_sort else None
+    key_columns = [*primary_key_columns(table), *([cursor] if cursor else [])]
+    if invalid := [column for column in key_columns if "," in column]:
+        raise DestinationTerminalException(
+            f"Table {table['name']}: column names {invalid} contain a comma, which the "
+            "comma-separated primary_key and cursor_field parameters cannot express."
+        )
+    params = {"primary_key": ",".join(primary_key_columns(table))}
+    if cursor:
+        params["cursor_field"] = cursor
+    return params
+
+
+def upload_mode(table: TTableSchema, table_already_replaced: bool) -> str:
+    if table.get("write_disposition") == "replace" and not table_already_replaced:
+        return "overwrite"
+    return "append"
+
+
+def replaced_tables() -> list[str]:
+    return dlt.current.destination_state().setdefault("replaced_tables", [])
+
+
+def evolved_tables() -> list[str]:
+    return dlt.current.destination_state().setdefault("evolved_tables", [])
+
+
+# batch_size=0 hands each load job a file path, the one branch of dlt's `TDataItems | str` argument.
+@dlt.destination(  # ty: ignore[invalid-argument-type]
+    name="altertable",
+    naming_convention="direct",
+    loader_file_format="parquet",
+    batch_size=0,
+    skip_dlt_columns_and_tables=False,
+    max_table_nesting=0,
+    loader_parallelism_strategy="table-sequential",
+    spec=AltertableClientConfiguration,
+)
+def altertable(
+    parquet_file_path: str,
+    table: TTableSchema,
+    config: AltertableClientConfiguration = dlt.config.value,
+) -> None:
+    upsert = upsert_params(table)
+    table_name = cast(str, table["name"])
+
+    already_replaced = replaced_tables()
+    already_evolved = evolved_tables()
+    mode = upload_mode(table, table_name in already_replaced)
+
+    if mode != "overwrite" and table_name not in already_evolved:
+        if create_or_evolve_table(config, table):
+            already_evolved.append(table_name)
+
+    params = {
+        "catalog": cast(str, config.catalog),
+        "schema": cast(str, config.dataset_name),
+        "table": table_name,
+    }
+    action = f"loading {config.catalog}.{config.dataset_name}.{table_name}"
+    if upsert is not None:
+        post_parquet(config, "upsert", params | upsert, parquet_file_path, action)
+    else:
+        params["mode"] = mode
+        with aligned_parquet(parquet_file_path, table) as upload_path:
+            post_parquet(config, "upload", params, upload_path, action)
+
+    if mode == "overwrite":
+        already_replaced.append(table_name)

dlt_altertable-0.1.0/src/dlt_altertable/table_schema.py

@@ -0,0 +1,126 @@
+import os
+import tempfile
+from collections.abc import Iterator
+from contextlib import contextmanager
+from typing import cast
+
+import pyarrow.parquet as pq
+from dlt.common import logger
+from dlt.common.data_writers.escape import escape_duckdb_literal, escape_postgres_identifier
+from dlt.common.destination.capabilities import DestinationCapabilitiesContext
+from dlt.common.exceptions import TerminalValueError
+from dlt.common.libs.pyarrow import normalize_py_arrow_item
+from dlt.common.normalizers.naming.direct import NamingConvention
+from dlt.common.schema import TTableSchema
+from dlt.common.schema.typing import TColumnSchema
+
+from dlt_altertable.api import execute_sql
+from dlt_altertable.configuration import AltertableClientConfiguration
+
+SQL_TYPES = {
+    "text": "VARCHAR",
+    "bigint": "BIGINT",
+    "double": "DOUBLE",
+    "bool": "BOOLEAN",
+    "timestamp": "TIMESTAMP WITH TIME ZONE",
+    "date": "DATE",
+    "time": "TIME",
+    "json": "VARCHAR",
+    "binary": "BLOB",
+    "decimal": "DECIMAL(38,9)",
+}
+
+
+def sql_type(column: TColumnSchema) -> str:
+    if column["data_type"] == "wei":
+        raise TerminalValueError(
+            f"Column {column['name']} has type wei, which needs 78 digits while DuckDB "
+            "decimals stop at 38."
+        )
+    if column["data_type"] == "decimal" and (precision := column.get("precision")) is not None:
+        return f"DECIMAL({precision},{column.get('scale', 0)})"
+    if column["data_type"] == "timestamp" and column.get("timezone") is False:
+        return "TIMESTAMP"
+    return SQL_TYPES[cast(str, column["data_type"])]
+
+
+def qualified_table_name(config: AltertableClientConfiguration, table_name: str) -> str:
+    return ".".join(
+        escape_postgres_identifier(part)
+        for part in cast(tuple[str, str, str], (config.catalog, config.dataset_name, table_name))
+    )
+
+
+def create_table(config: AltertableClientConfiguration, table: TTableSchema) -> None:
+    table_name = cast(str, table["name"])
+    columns = ", ".join(
+        f"{escape_postgres_identifier(name)} {sql_type(column)}"
+        for name, column in table["columns"].items()
+    )
+    execute_sql(
+        config,
+        "CREATE SCHEMA IF NOT EXISTS "
+        f"{escape_postgres_identifier(cast(str, config.catalog))}"
+        f".{escape_postgres_identifier(cast(str, config.dataset_name))}",
+    )
+    execute_sql(
+        config,
+        f"CREATE TABLE IF NOT EXISTS {qualified_table_name(config, table_name)} ({columns})",
+    )
+    logger.info(
+        f"Created table {config.catalog}.{config.dataset_name}.{table_name} "
+        f"with {len(table['columns'])} columns"
+    )
+
+
+def create_or_evolve_table(config: AltertableClientConfiguration, table: TTableSchema) -> bool:
+    """Creating the table from dlt's typed schema keeps column types and later evolution
+    deliberate instead of whatever the server would infer from the first file. A table left by
+    an earlier load must also gain the columns that dlt's schema evolution added since.
+
+    Returns whether the lookup found an existing table, the only answer worth caching for the
+    rest of the load: a CREATE is never read back to confirm what the table now holds."""
+    table_name = cast(str, table["name"])
+    rows = execute_sql(
+        config,
+        "SELECT column_name FROM information_schema.columns "
+        f"WHERE table_catalog = {escape_duckdb_literal(config.catalog)} "
+        f"AND table_schema = {escape_duckdb_literal(config.dataset_name)} "
+        f"AND table_name = {escape_duckdb_literal(table_name)}",
+    )
+    existing = {row[0] for row in rows}
+    if not existing:
+        create_table(config, table)
+        return False
+    for name, column in table["columns"].items():
+        if name not in existing:
+            execute_sql(
+                config,
+                f"ALTER TABLE {qualified_table_name(config, table_name)} "
+                f"ADD COLUMN IF NOT EXISTS {escape_postgres_identifier(name)} {sql_type(column)}",
+            )
+            logger.info(
+                f"Added column {name} to {config.catalog}.{config.dataset_name}.{table_name}"
+            )
+    return True
+
+
+@contextmanager
+def aligned_parquet(parquet_file_path: str, table: TTableSchema) -> Iterator[str]:
+    """dlt can evolve a load's schema between files, leaving an earlier file narrower than the
+    table the load builds, while an append or replace upload must match the table column for
+    column. A file that falls short is padded with typed NULL columns into a temporary copy.
+    Merge files never come through here, so columns a merge file omits keep their stored values."""
+    if pq.read_schema(parquet_file_path).names == list(table["columns"]):
+        yield parquet_file_path
+        return
+    aligned = normalize_py_arrow_item(
+        pq.read_table(parquet_file_path),
+        table["columns"],
+        NamingConvention(),
+        DestinationCapabilitiesContext.generic_capabilities(),
+    )
+    with tempfile.TemporaryDirectory() as aligned_dir:
+        aligned_path = os.path.join(aligned_dir, "aligned.parquet")
+        pq.write_table(aligned, aligned_path)
+        yield aligned_path