Cartage

Declarative data migrations. Describe sources, destinations and pipelines in YAML, put custom logic in plain Python, run locally to test, and generate thin Airflow DAGs, Dagster jobs or Prefect flows for production.

Cartage is an abstraction and simplification layer over data load tools. You describe what to move and how to reshape it; an engine moves the data. dlt is the default engine. The engine is pluggable, so other load tools can be supported later without changing your pipelines.

  what you write      cartage.yaml · connections.yaml · pipelines/*.yaml · transforms/*.py
                                              │
                                              ▼
                ┌───────────────────────── Cartage ─────────────────────────┐
                │  validate · plan · run · state · rejects                  │
                │  secrets: ${secret:…} ${env:…} ${airflow:…}               │
                │  generate: Airflow · Dagster · Prefect  (they call `run`) │
                └─────────────────────────────┬─────────────────────────────┘
                                              │ runs each pipeline on
                ┌───────────────────────── engine ──────────────────────────┐
                │  dlt (default)  │  python (reference)  │  future: others  │
                └───────┬─────────────────────────────────────────┬─────────┘
                        │ reads                                   │ writes
  files (local · S3 · GCS · HTTP) · dlt sources   SAP BAPIs · file exports · dlt destinations
                        └──── your Python transforms in between ──┘

• Sources: files (CSV, JSON Lines, Parquet) in a local folder, S3, GCS, Azure or over HTTP, and any dlt source (REST APIs, databases, SaaS) from YAML or Python.
• Destinations: any dlt destination (Snowflake, DuckDB, Postgres, BigQuery, a filesystem data lake, ...), loaded natively by dlt; SAP via BAPIs (ships a mock SAP; the RFC transport is planned); and file exports (JSON, JSON Lines, XML, CSV). One pipeline can write to several destinations.
• Connections are named after the system (type: snowflake, type: filesystem, type: sap_bapi), never after the engine: see docs/connections.md.
• Engines: dlt (default) and python, a plain-loop reference engine for tests (see Engines).
• Orchestrators: Airflow, Dagster and Prefect. Generated code only calls cartage run, so local and production run the same code.

Quickstart

pip install cartage
cartage init demo --answers https://raw.githubusercontent.com/datacoves/cartage/main/examples/sap/answers.yaml --yes
cd demo
cartage validate
cartage plan materials        # dry run: records before/after transforms and the BAPI payloads
cartage run materials         # reads 20 rows, filters 2, sends 18: 16 load, 2 fail on purpose (exit 1)
cartage run materials --advance-state
cartage run materials         # incremental: nothing new to load
cartage generate              # dags/materials.py

The demo runs in dev on a sample file and a mock SAP; its prd environment reads S3 and calls SAP, once the "<fill me>" placeholders are filled. Run cartage init <dir> without --answers to create your own project; see Creating a project.

Split-screen demo: run cartage sap mock in one terminal, set url: http://localhost:8765 on sap.dev in connections.yaml, and run the pipeline in another terminal.

Creating a project

cartage init <dir> asks a few questions and writes a project that validates straight away:

1. the project name and environments (dev, prd; the first is the default),
2. the source system: files (a folder, s3://, gs://, az://, https://), a REST API, a SQL database or a Python function, plus a few details for it,
3. the destination: any dlt destination (snowflake, duckdb, postgres, ...), file_export or sap_bapi,
4. an optional sample file: with one, the first environment is fully local (the sample replaces the source through source.envs, and the destination becomes DuckDB, the SAP mock or a local folder), so cartage run works before any credentials exist,
5. the pipeline name and an optional schedule (Airflow, Dagster or Prefect).

Credentials become placeholders: settings in connections.yaml and ${secret:<connection>.<env>.<field>} references to .cartage/secrets.yaml, both "<fill me>". The fields come from dlt's own credential classes, so every dlt destination is covered; optional fields are listed as comments. cartage validate --env <env> names every placeholder still to fill.

--answers <path-or-url> answers questions from a file (the rest are asked, or take their defaults with --yes). An answers file can also carry pipeline content: transforms, destination_options and copy (files to copy, relative to the answers file). examples/sap/answers.yaml builds the demo above this way.

Install

Cartage needs Python 3.11.4 or newer. dlt is included; extras add the rest.

cartage[dlt] still works; since 0.6 it adds nothing, as dlt is a core dependency.

Generating Airflow, Dagster or Prefect files needs nothing extra; the orchestrator itself runs them.

Engines

The engine runs a pipeline: it reads from the source, applies your transforms, and hands records to the destination. Set it in cartage.yaml (defaults.engine), per pipeline (engine:), or per run (--engine).

• dlt (default) runs every pipeline as a dlt pipeline: dlt's extraction, normalization, incremental state and loading. dlt destinations load natively (bulk loads, staging, merge); record destinations (sap_bapi, file_export) plug in as a dlt custom destination, so Cartage can still report each record's result (rejects, warnings).
• python is a plain loop: the reference implementation of the engine contract, for tests and framework development. It reads sources and writes record destinations; dlt destinations and incremental sources need the dlt engine, which keeps the cursor.

The engines differ in two ways today: dlt drops fields whose value is None before a record destination sees them, and the dlt engine hands record destinations batches of 100 (which is what SAP commit: per_batch groups).

Engines are plug-ins (the cartage.engines entry point group, see Extending), so another load tool can become an engine without changing pipelines, as long as it can run Python transforms between reading and writing.

Commands

Global options go before the command: -C/--project-dir, -v/--verbose, --debug (tracebacks), --no-color.

Project layout

YAML configuration

Cartage uses three YAML layers: cartage.yaml sets project-wide defaults, connections.yaml defines named services per environment, and each pipelines/*.yaml file describes one flow from a source to one or more destinations.

cartage.yaml

project: inventory
environments: [dev, prd]
default_env: dev

defaults:
  engine: dlt # or python

state:
  dev: { path: .cartage/state }

orchestrators:
  airflow:
    dags_dir: dags
    default_args: { owner: data-team, retries: 1 }

The default_env must be listed in environments. A pipeline can override the default engine with its own engine. State settings are optional; when omitted, Cartage stores local state under .cartage/state.

connections.yaml

Connections have a type and an envs map. Put service-specific settings under the environment where they apply; pipelines refer to the connection by name. Keep credentials out of the file and use secret or environment references.

connections:
  local_files:
    type: filesystem                  # a local folder, s3://, gs://, az://, https://, ...
    envs:
      dev: { bucket_url: ./data }

  warehouse:
    type: snowflake                   # any dlt destination: duckdb, postgres, bigquery, ...
    envs:
      dev: { type: duckdb, credentials: dev.duckdb }     # an environment may use another system
      prd: { credentials: "${secret:snowflake.connection_string}" }

  sap:
    type: sap_bapi
    envs:
      dev: { transport: mock, client: "100" }
      prd:
        transport: rfc
        ashost: sap.example.com
        sysnr: "00"
        client: "100"
        user: "${secret:sap.user}"
        passwd: "${secret:sap.passwd}"

${secret:key} resolves from CARTAGE_SECRET__<KEY> (dots become double underscores and names are uppercased), then .cartage/secrets.yaml. ${env:NAME} reads an environment variable directly. See Secrets for details.

docs/connections.md lists every connection type (filesystem, the dlt destinations, file_export, sap_bapi): which settings go in connections.yaml and which options go in the pipeline.

pipelines/*.yaml

Each pipeline names a source connection, applies an ordered list of transforms, and writes to a destination (or a list of destinations, see Multiple destinations). A step must have exactly one of map, filter, or batch; each reference uses the module:function format.

name: materials
source:
  connection: local_files
  format: csv
  path: materials/*.csv
  incremental: true

transforms:
  - map: transforms.materials:normalize_uom
  - filter: transforms.materials:is_active
  - batch: transforms.materials:dedupe
    with: { key: material }

destination:
  connection: sap
  bapi: BAPI_MATERIAL_SAVEDATA
  mapping:
    HEADDATA.MATERIAL: material
    MATERIALDESCRIPTION[].MATL_DESC: description
    CLIENTDATA.BASE_UOM: uom
  constants:
    MATERIALDESCRIPTION[].LANGU_ISO: EN
  commit: per_record

schedule:
  airflow:
    schedule: "0 3 * * *"
    tags: [sap, materials]

Transforms run top to bottom, and each kind can appear any number of times and in any order, so several maps, filters or batches chain naturally (each step with its own with:):

transforms:
  - map: transforms.materials:normalize_uom
  - map: transforms.materials:trim_text
  - filter: transforms.materials:is_active
  - map: transforms.materials:add_defaults
    with: { plant: "1000" }
  - filter: transforms.materials:has_description
  - batch: transforms.materials:dedupe
    with: { key: material }

Source and destination fields other than connection are adapter options. Transform functions live in your transforms/ package; with passes keyword arguments to the function, and on_error: fail stops the run on that step's first error instead of rejecting the record. The optional schedule.airflow, schedule.dagster and schedule.prefect blocks control generation with cartage generate (see Orchestrators). Run cartage validate after editing YAML to check the project and references.

Sources

Full reference for every type: docs/connections.md.

A filesystem source reads every file matching path with dlt's typed readers (pandas for CSV), so numbers arrive as numbers. reader_options go to the reader, e.g. { dtype: str, keep_default_na: false } to keep codes such as 000123 as text and empty cells as "". With incremental: true only files modified since the last run are read (a changed file is read again in full). The same connection can be a destination (a dlt data lake) and the state store for SAP and file export pipelines.

A dlt source's ref returns a dlt source or resource; incremental: { cursor: updated_at, initial: "2024-01-01" } adds a dlt cursor. Incremental sources need the dlt engine (the default). Other keys go to dlt.sources.incremental: lag (re-read a window before the last value, e.g. for late updates), end_value, primary_key, row_order, last_value_func, on_cursor_value_missing, range_start, range_end. When ref returns a dlt source with several resources, resources: [name, ...] picks the ones to read (default: the source's selected resources).

Generic sources (no Python)

Files need no code: point a filesystem connection at a folder, a bucket or a web server (bucket_url: https://raw.githubusercontent.com/datasets/population/master/data, then path: population.csv). ref also takes installed modules, so dlt's built-in sources need only YAML, e.g. a REST API:

# dlt's declarative rest_api source; `config` is its configuration dict.
source:
  ref: dlt.sources.rest_api:rest_api_source
  with:
    config:
      client: { base_url: "https://earthquake.usgs.gov/fdsnws/event/1/" }
      resources:
        - name: earthquakes
          endpoint:
            path: query
            params: { format: geojson, starttime: "2026-10-01", endtime: "2026-10-02" }
            data_selector: features      # where the records are in the response
            paginator: single_page

Remote files need the matching dlt extra: dlt[http] for https://, dlt[s3], dlt[gs], dlt[az]. CSV and JSON Lines are read as records, so a column that is empty in every row is not created; declare it in the destination's columns ({ my_column: { data_type: double } }) if it must exist. See dlt's REST API docs for every option (authentication, pagination, incremental parameters). Reshape records with transforms; write a Python source only for logic the configuration can't express.

SAP destination (sap_bapi)

bapi names the BAPI, mapping fills BAPI parameters (HEADDATA.MATERIAL, TABLE[].FIELD) from record fields and constants fills them with fixed values; both are keyed by the BAPI parameter. commit is per_record (default), per_batch or none. cartage validate checks mapped parameters and fields against the bundled BAPI metadata; the mock SAP also rejects values that are too long, not allowed, or missing when required, as SAP would.

File export destination (file_export)

type: file_export writes each run to one local file, in json (an array), jsonl (default), xml or csv. It works with both engines and needs no extra dependencies. cartage plan shows each record as it will be written.

connections:
  exports:
    type: file_export
    envs:
      dev: { path: ./output }       # folder, relative to the project

destinations:
  - { connection: exports, format: json }                                 # output/<pipeline>.json
  - { connection: exports, name: as_xml, format: xml, root: materials, record: material }
  - { connection: exports, name: as_csv, format: csv, file: "materials_{date}.csv" }

file is the name inside the folder ({pipeline}, {date} as YYYYMMDD UTC, {ext}; default {pipeline}.{ext}). The file is written to <file>.partial and moved into place only when the run succeeds, so a failed run never leaves a half-written file and the previous one stays. Nested values become nested elements in XML (lists as <item>) and JSON text in CSV; XML names that start with a digit get a leading _. A CSV's columns come from its first record; records with other columns are rejected. Records that fail transforms go to rejects as usual, and the file holds the rest.

dlt destinations

Every dlt destination is a connection type of its own: snowflake, duckdb, postgres, bigquery, databricks, filesystem (a data lake), ... (cartage plugins lists them). The settings are that destination's dlt settings (credentials, database, bucket_url, ...). A project module:function returning a dlt destination works as a type too. Install the driver extra, e.g. pip install "dlt[snowflake]". dlt destinations need the dlt engine (the default).

connections:
  warehouse:
    type: snowflake
    envs:
      prd:
        credentials: "${secret:snowflake.connection_string}"
        naming: sql_ci_v1                              # connection-level default
  lake:
    type: filesystem                                   # also a source and a state store for sinks
    envs:
      dev: { bucket_url: ./lake }

destination:
  connection: warehouse
  dataset_name: raw                 # default: the pipeline name
  table_name: materials             # default: the source resource name
  write_disposition: merge          # append | replace | merge; unset: the source resource's hint, else append
  primary_key: material             # merge key: here, or the source resource's own hints
  columns: { price: { data_type: decimal, precision: 18, scale: 4 } }
  loader_file_format: parquet
  naming: direct                    # dlt naming convention, or a project module (e.g. naming.upper)
  dlt_config:                       # any dlt config key, applied to this run only
    data_writer.buffer_max_items: 100000
    extract.max_parallel_items: 15
  after_load:                       # called as f(pipeline, **with) after a successful load
    - { ref: utils.snowflake:enable_change_tracking, with: { tables: [materials] } }

The dlt engine loads straight into the destination: bulk loads, staging and merge are dlt's, run once per run. Without transforms, source batches reach dlt untouched, so Arrow/pandas batches (e.g. sql_database with the pyarrow backend, ConnectorX) stay columnar. Hints the source resource sets itself (write_disposition, primary_key, columns, ...) are kept unless the destination sets them. after_load hooks get the dlt pipeline, so they can run SQL with pipeline.sql_client() (grants, tags, change tracking); if one fails the data stays loaded, the run exits with code 3 and state is not saved. dlt fails whole load jobs, not single records: a failed load stops the run and state is not saved. Database-specific extraction (casts for Oracle LOBs and NUMBERs, MSSQL datetime2, Postgres enums and timestamptz, time zones) belongs in your dlt source factory, not in Cartage.

Example: an incremental table copy from a database into Snowflake, kept columnar end to end.

# sources/erp.py
from dlt.sources.sql_database import sql_table

def materials(credentials: str, schema: str):
    return sql_table(credentials=credentials, schema=schema, table="materials", backend="pyarrow")

name: erp_materials
source:
  ref: sources.erp:materials
  with: { credentials: "${secret:erp.url}", schema: ERP }
  incremental: { cursor: updated_at, initial: "2024-01-01" }
destination:
  connection: warehouse
  dataset_name: raw_erp
  write_disposition: merge
  primary_key: material
engine: dlt

Multiple destinations

Use destinations: instead of destination:. Each entry is its own run, in order, with its own state, rejects and dlt pipeline (<pipeline>__<name>; name defaults to the connection and tells apart destinations that share one), so a failed destination retries without reloading the others. The source is read once per destination; to read it once, land it in a filesystem connection first and load from there.

destinations:
  - { connection: warehouse, table_name: materials, write_disposition: merge, primary_key: material }
  - { connection: sap, bapi: BAPI_MATERIAL_SAVEDATA, mapping: { ... } }

Orchestrators

cartage generate --target airflow|dagster|prefect writes one file per pipeline that has a schedule.<target> block. Each file only runs cartage --project-dir <project> run <pipeline> --env <env>, so deploy the Cartage project (with transforms/) next to it. Settings merge: built-in defaults ← orchestrators.<target> in cartage.yaml ← the pipeline's schedule.<target> (dicts merge by key). Add --check in CI to fail on stale files.

schedule:
  airflow: { schedule: "0 3 * * *", tags: [sap] }
  dagster: { schedule: "0 3 * * *", tags: { team: data }, timezone: Europe/Brussels }
  prefect: { schedule: "0 3 * * *", tags: [sap], retries: 2 }

• Dagster: each file defines a job, its schedule, and defs = dg.Definitions(...). Load it with dagster dev -f orchestration/dagster/<name>.py (repeat -f per pipeline), or merge the defs objects.
• Prefect: each file defines the cartage_run flow. python orchestration/prefect/<name>.py serves it on the schedule, or point prefect deploy at orchestration/prefect/<name>.py:cartage_run.
• env defaults to prd; command (default cartage) may include a launcher, e.g. uv run cartage.
• Airflow: DAGs use the TaskFlow API (@dag, @task.bash). task_decorator picks the task decorator, e.g. datacoves_bash for @task.datacoves_bash; operator_args are its keyword arguments. task_env adds environment variables to the task (e.g. UV_CACHE_DIR). See Airflow settings for packages, owner/email and functions that build default_args or the schedule.

Airflow settings

# cartage.yaml
orchestrators:
  airflow:
    task_decorator: datacoves_bash
    command: uvx --from "cartage>=0.8.0" cartage
    dependencies: ["dlt[snowflake,parquet]"]                  # every DAG
    default_args: { owner: data-team, email: [data@example.com], retries: 1 }

# pipelines/loans.yaml
schedule:
  airflow:
    schedule: "0 3 * * *"
    dependencies: ["dlt[http]"]                               # only this DAG
    default_args: { owner: Noel Gomez, email: [noel@example.com] }

• dependencies from cartage.yaml and the pipeline add up (every other list replaces) and become --with flags right after a uvx, uv tool run or uv run command: uvx --with 'dlt[snowflake,parquet]' --with 'dlt[http]' --from 'cartage>=0.8.0' cartage.

• default_args merge by key, so the owner and email can be set once and overridden per pipeline.

• Values are written into the DAG as Python literals. When default_args need Python (timedeltas, callbacks), point default_args_from at a function that builds them; default_args become its keyword arguments. schedule_from wraps the schedule the same way, e.g. to turn it off outside production:

  default_args_from: orchestrate.utils.datacoves_utils:set_default_args
  default_args: { owner: Noel Gomez, owner_email: noel@example.com }
  schedule_from: orchestrate.utils.datacoves_utils:set_schedule
  

  renders default_args = set_default_args(owner='Noel Gomez', owner_email='noel@example.com') and schedule=set_schedule('0 3 * * *').

• For anything else, a project template (templates/airflow/dag.py.j2, {% extends "cartage/airflow_dag.py.j2" %}) can override the blocks header, imports, default_args, schedule, dag, task and extra.

Airflow connections

In Airflow, credentials usually live in Airflow connections. Cartage reads them with ${airflow:...} references, so nothing has to copy them into variables by hand and they never appear in YAML or in the generated DAG.

1. Reference connection fields in connections.yaml, in the environment the DAGs run (the cartage_env Airflow variable, prd by default):

warehouse:
  type: snowflake
  envs:
    dev: { destination_name: datacoves_snowflake }   # credentials from ~/.dlt/secrets.toml locally
    prd:
      credentials:
        host: "${airflow:main_load_keypair.extra.account}"
        username: "${airflow:main_load_keypair.login}"
        private_key: "${airflow:main_load_keypair.extra.private_key_content}"   # PEM, as stored
        warehouse: "${airflow:main_load_keypair.extra.warehouse}"
        role: "${airflow:main_load_keypair.extra.role}"

2. Generate the DAGs. cartage generate collects every ${airflow:...} used by the pipeline (its source, destinations, state store and options, in every environment) and passes them to the task as environment variables whose values are Airflow templates:

@task.bash(
    env={"CARTAGE_AIRFLOW__MAIN_LOAD_KEYPAIR__LOGIN": "{{ conn.get('main_load_keypair').login or '' }}", ...},
    append_env=True,
)
def cartage_run():
    return "cartage --project-dir ... run pipelines/loans.yaml --env {{ var.value.get('cartage_env', 'prd') }}"

3. Airflow fills them in when the task runs, from wherever it keeps connections (metadata database, a secrets backend such as Vault or AWS Secrets Manager, or AIRFLOW_CONN_<ID> variables), and masks passwords in the task log.

4. cartage run resolves each reference from its CARTAGE_AIRFLOW__<CONN_ID>__<FIELD> variable (dots and dashes become __ and _). Values are never printed.

Notes:

• Rerun cartage generate when references change; cartage generate --check in CI catches stale DAGs.
• A field that is empty in the connection resolves to "". A missing extra key fails the task when Airflow renders it (Airflow's StrictUndefined), naming the key.
• Outside Airflow, cartage run/validate stop at the first ${airflow:...} with a hint naming the variable. Use another environment locally (${secret:...}, ${env:...} or the driver's own config), or export the variable.
• To test a DAG without a real connection, define one inline: AIRFLOW_CONN_MAIN_LOAD_KEYPAIR='{"conn_type": "snowflake", "login": "...", "extra": {...}}' airflow dags test <dag_id>.
• Dagster and Prefect pass secrets to tasks as environment variables; read them with ${env:...}.

Batches

• batch transforms see one source batch at a time (default 100 rows, batch_size source option).
• commit: per_batch commits each batch the engine hands to the destination (the dlt engine re-chunks at 100).
• dlt destinations are not fed in batches: dlt extracts, normalizes and loads the whole run. Size its files and buffers with dlt_config (e.g. data_writer.buffer_max_items, data_writer.file_max_items).

Secrets

${secret:sap.passwd} reads CARTAGE_SECRET__SAP__PASSWD, then .cartage/secrets.yaml (git-ignored; the dots are nesting levels). ${env:VAR} reads an environment variable. Resolved values are never printed.

# .cartage/secrets.yaml
sap: { user: rfc_user, passwd: "..." }
snowflake: { connection_string: "snowflake://user:...@account/db?warehouse=wh&role=r" }

dlt destinations can also leave credentials out of connections.yaml entirely: dlt then reads its own .dlt/secrets.toml (in the working directory) or ~/.dlt/secrets.toml (e.g. destination_name: my_snowflake reads [destination.my_snowflake.credentials]).

${airflow:<conn_id>.<field>} reads a field of an Airflow connection inside DAGs from cartage generate; see Airflow connections.

State and rejects

dlt destinations (snowflake, duckdb, postgres, filesystem, ...) keep their state the way any dlt pipeline does: in the destination (_dlt_pipeline_state), loaded together with the data and restored from there on the next run, so a fresh Airflow worker needs no state settings. Record errors don't hold that state back: the good rows and the new cursor land together, and rejected records stay in the rejects file. --full-refresh drops the pipeline's tables and state, then loads again.

dlt cannot keep state in SAP or in a file export, so those pipelines use Cartage's state store. It is saved only when a run has no record errors (or with --advance-state), and it lives in .cartage/state unless cartage.yaml points an environment at a filesystem connection: state: { prd: { connection: landing, prefix: cartage/state/ } } (S3, GCS, Azure, ...). Local state is lost on orchestrator workers, so cartage generate refuses a SAP or file export pipeline whose environment keeps it on local disk. --full-refresh ignores the stored state.

Rejected records go to .cartage/rejects/<pipeline>/<run_id>.jsonl. cartage state show|reset <pipeline> inspects or clears state, wherever it lives.

With several destinations, each destination is its own run named <pipeline>__<name>: state, rejects and the dlt pipeline are kept per destination, and state show|reset covers all of them. cartage run goes through the destinations in order and stops at the first fatal error; destinations that already finished keep their state, so a rerun only retries the rest. With --json, run prints one result line per destination and plan prints a list.

Exit codes

0 ok · 1 record errors · 2 configuration error · 3 fatal run error (connection, transport, failed dlt load, on_error: fail).

Advanced: database extraction tuning

Cartage has no database drivers of its own: a dlt source's ref is your Python, so ConnectorX, sql_table, SQLAlchemy, Oracle or Teradata clients, and every per-database fix are plain code in your project. Cartage calls the function with with: (secrets resolved), adds incremental:, and hands what it yields to dlt. Keep tuning helpers in a project package (e.g. utils/) and use them from source functions. Where each kind of tuning goes:

Driver, casts and fallback

The source function picks the driver and can fall back when a type isn't supported, then build its query from metadata-driven casts (for example a build_select_with_casts-style helper that reads ALL_TAB_COLUMNS or INFORMATION_SCHEMA):

# sources/erp.py
from urllib.parse import quote_plus

import connectorx as cx
import dlt

from utils.data_types import build_select  # your helper, e.g. adapted from build_select_with_casts


@dlt.resource(name="materials", max_table_nesting=0)
def materials(db: str, user: str, password: str, dsn: str, table: str, conversions: dict | None = None,
              updated_at=dlt.sources.incremental("UPDATED_AT")):
    conn = f"{db}://{quote_plus(user)}:{quote_plus(password)}@{dsn}"
    where = f" WHERE UPDATED_AT >= '{updated_at.start_value}'" if updated_at.start_value else ""
    query = f"SELECT {build_select(conn, table, conversions or {})} FROM {table}{where}"
    try:
        yield from cx.read_sql(conn, query, return_type="arrow_stream", batch_size=10_000)
    except BaseException:  # Arrow stream lacks some types (e.g. Oracle NCLOB): fall back to a DataFrame
        yield cx.read_sql(conn, query)

source:
  ref: sources.erp:materials
  with:
    db: oracle
    user: "${secret:erp.user}"
    password: "${secret:erp.password}"
    dsn: "${secret:erp.dsn}"
    table: ERP.MATERIALS
    conversions: { CREATED_ON: timestamp, NOTES: lob_text }
  incremental: { cursor: UPDATED_AT, initial: "2024-01-01" }

Cartage's incremental: binds to the function's dlt.sources.incremental argument, so start_value is the stored cursor: put it in the WHERE to read only new rows. dlt's sql_table(..., backend="connectorx" | "pyarrow" | "sqlalchemy") does that pushdown for you, and takes query_adapter_callback (extra filters or suffixes), table_adapter_callback (e.g. remove_nullability_adapter) and chunk_size. Pass with: values such as backend or connector: sqlalchemy to switch drivers per pipeline when ConnectorX can't read a table.

Per-batch type normalization

Attach Arrow, pandas or row transforms with add_map. They run inside dlt on every batch, before normalization, and keep batches columnar:

from utils.data_types import normalize_types  # timestamps → epoch ms, decimals/ints → float64, rounding

def orders(...):
    return sql_table(credentials=..., table="ORDERS", backend="pyarrow").add_map(normalize_types)

Prefer add_map in the source over Cartage transforms for bulk copies: Cartage transforms see row dicts, so Arrow batches are converted to rows first.

Row-count checks

ConnectorX can silently return fewer rows (dropped batches, timeouts). Count what you yield and raise at the end; the run then fails with exit code 3 before anything is loaded, and no state is saved:

def checked(batches, expected: int, tolerance: int = 10):
    n = 0
    for batch in batches:
        n += len(batch)  # Arrow tables, DataFrames and lists all have len()
        yield batch
    if n < expected - tolerance:
        raise RuntimeError(f"extracted {n} rows, source had {expected}: try connector: sqlalchemy")

@dlt.resource(name="materials")
def materials(conn: str, table: str):
    expected = int(cx.read_sql(conn, f"SELECT COUNT(*) FROM {table}").iloc[0, 0])
    yield from checked(cx.read_sql(conn, f"SELECT * FROM {table}", return_type="arrow_stream"), expected)

dlt settings and naming

dlt_config: takes any dlt config key (as in dlt.config[...]), on the connection as a default or on the pipeline destination as an override, applied only while that run executes:

destination:
  connection: lake
  loader_file_format: parquet
  write_disposition: replace
  naming: utils.sql_upper             # a module with a NamingConvention class, as dlt's schema.naming expects
  columns: { AMOUNT: { data_type: double, nullable: true } }
  dlt_config:
    sources.data_writer.file_max_items: 1000000
    sources.data_writer.file_max_bytes: 100000000
    data_writer.buffer_max_items: 100000
    extract.max_parallel_items: 15
    normalize.data_writer.disable_compression: false
    load.truncate_staging_dataset: true

naming takes dlt's built-ins (direct, snake_case, sql_cs_v1, sql_ci_v1) or a project module, so naming classes such as upper, lower or keep-quotes variants work as they are. Two settings are Cartage's: leave restore_from_destination on (the default; it is how a run gets its state back from the destination) and load.delete_completed_jobs true (the SAP and file export state archive must not keep loaded data).

Process-wide patches to dlt internals (for example keeping all-null columns) are not run settings: apply them in the source module (at import, or at the start of the source function), knowing they affect every pipeline in that process.

Extending

Adapters are entry points in the groups cartage.sources, cartage.destinations, cartage.engines and cartage.orchestrators. cartage plugins lists what is installed.

A destination is a record sink (preview + write, like sap_bapi and file_export) by default. Optional: finish(ok) is called after the run (to finalize or discard output), and preview_label/preview_syntax name and highlight plan output. To have the dlt engine load into it natively instead, expose dlt_destination() returning a dlt destination, plus hints (resource hints such as write_disposition), dataset_name, loader_file_format and dlt_env (env vars applied to the run), as the dlt destination adapter does. dlt destination names (snowflake, filesystem, ...) are reserved: a plug-in with one of those names is an error.

License

Apache-2.0