Schema

The schema describes the structure of normalized data (e.g. tables, columns, data types, etc.) and provides instructions on how the data should be processed and loaded. dlt generates schemas from the data during the normalization process. User can affect this standard behavior by providing hints that change how tables, columns and other metadata is generated and how the data is loaded. Such hints can be passed in the code ie. to dlt.resource decorator or pipeline.run method. Schemas can be also exported and imported as files, which can be directly modified.

💡 dlt associates a schema with a source and a table schema with a resource.

Schema content hash and version

Each schema file contains content based hash version_hash that is used to:

1. Detect manual changes to schema (ie. user edits content).
2. Detect if the destination database schema is synchronized with the file schema.

Each time the schema is saved, the version hash is updated.

Each schema contains a numeric version which increases automatically whenever schema is updated and saved. Numeric version is meant to be human-readable. There are cases (parallel processing) where the order is lost.

💡 Schema in the destination is migrated if its hash is not stored in _dlt_versions table. In principle many pipelines may send data to a single dataset. If table name clash then a single table with the union of the columns will be created. If columns clash, and they have different types etc. then the load may fail if the data cannot be coerced.

Naming convention

dlt creates tables, child tables and column schemas from the data. The data being loaded, typically JSON documents, contains identifiers (i.e. key names in a dictionary) with any Unicode characters, any lengths and naming styles. On the other hand the destinations accept very strict namespaces for their identifiers. Like Redshift that accepts case-insensitive alphanumeric identifiers with maximum 127 characters.

Each schema contains naming convention that tells dlt how to translate identifiers to the namespace that the destination understands.

The default naming convention:

1. Converts identifiers to snake_case, small caps. Removes all ascii characters except ascii alphanumerics and underscores.
2. Adds _ if name starts with number.
3. Multiples of _ are converted into single _.
4. The parent-child relation is expressed as double _ in names.
5. It shorts the identifier if it exceed the length at the destination.

💡 Standard behavior of dlt is to use the same naming convention for all destinations so users see always the same tables and columns in their databases.

💡 If you provide any schema elements that contain identifiers via decorators or arguments (i.e. table_name or columns) all the names used will be converted via the naming convention when adding to the schema. For example if you execute dlt.run(... table_name="CamelCase") the data will be loaded into camel_case.

💡 Use simple, short small caps identifiers for everything!

To retain the original naming convention (like keeping "createdAt" as it is instead of converting it to "created_at"), you can use the direct naming convention, in "config.toml" as follows:

[schema]
naming="direct"

caution

Opting for "direct" naming bypasses most name normalization processes. This means any unusual characters present will be carried over unchanged to database tables and columns. Please be aware of this behavior to avoid potential issues.

The naming convention is configurable and users can easily create their own conventions that i.e. pass all the identifiers unchanged if the destination accepts that (i.e. DuckDB).

Data normalizer

Data normalizer changes the structure of the input data, so it can be loaded into destination. The standard dlt normalizer creates a relational structure from Python dictionaries and lists. Elements of that structure: table and column definitions, are added to the schema.

The data normalizer is configurable and users can plug their own normalizers i.e. to handle the parent-child table linking differently or generate parquet-like data structs instead of child tables.

Tables and columns

The key components of a schema are tables and columns. You can find a dictionary of tables in tables key or via tables property of Schema object.

A table schema has the following properties:

1. name and description.
2. parent with a parent table name.
3. columns with dictionary of table schemas.
4. write_disposition hint telling dlt how new data coming to the table is loaded.

Table schema is extended by data normalizer. Standard data normalizer adds propagated columns to it.

A column schema contains following properties:

1. name and description of a column in a table.
2. data_type with a column data type.
3. precision a precision for text, timestamp, time, bigint, binary, and decimal types
4. scale a scale for decimal type
5. is_variant telling that column was generated as variant of another column.

A column schema contains following basic hints:

1. nullable tells if column is nullable or not.
2. primary_key marks a column as a part of primary key.
3. merge_key marks a column as a part of merge key used by incremental load.
4. foreign_key marks a column as a part of foreign key.
5. root_key marks a column as a part of root key which is a type of foreign key always referring to the root table.
6. unique tells that column is unique. on some destination that generates unique index.

dlt lets you define additional performance hints:

1. partition marks column to be used to partition data.
2. cluster marks column to be part to be used to cluster data
3. sort marks column as sortable/having order. on some destinations that non-unique generates index.

note

Each destination can interpret the hints in its own way. For example cluster hint is used by Redshift to define table distribution and by BigQuery to specify cluster column. DuckDB and Postgres ignore it when creating tables.

Variant columns

Variant columns are generated by a normalizer when it encounters data item with type that cannot be coerced in existing column. Please see our coerce_row if you are interested to see how internally it works.

Let's consider our getting started example with slightly different approach, where id is an integer type at the beginning

data = [
  {"id": 1, "human_name": "Alice"}
]

once pipeline runs we will have the following schema:

name	data_type	nullable
id	bigint	true
human_name	text	true

Now imagine the data has changed and id field also contains strings

data = [
  {"id": 1, "human_name": "Alice"},
  {"id": "idx-nr-456", "human_name": "Bob"}
]

So after you run the pipeline dlt will automatically infer type changes and will add a new field in the schema id__v_text to reflect that new data type for id so for any type which is not compatible with integer it will create a new field.

name	data_type	nullable
id	bigint	true
human_name	text	true
id__v_text	text	true

On the other hand if id field was already a string then introducing new data with id containing other types will not change schema because they can be coerced to string.

Now go ahead and try to add a new record where id is float number, you should see a new field id__v_double in the schema.

Data types

dlt Data Type	Source Value Example	Precision and Scale
text	'hello world'	Supports precision, typically mapping to VARCHAR(N)
double	45.678
bool	True
timestamp	'2023-07-26T14:45:00Z', datetime.datetime.now()	Supports precision expressed as parts of a second
date	datetime.date(2023, 7, 26)
time	'14:01:02', datetime.time(14, 1, 2)	Supports precision - see timestamp
bigint	9876543210	Supports precision as number of bits
binary	b'\x00\x01\x02\x03'	Supports precision, like text
complex	[4, 5, 6], {'a': 1}
decimal	Decimal('4.56')	Supports precision and scale
wei	2**56

wei is a datatype tries to best represent native Ethereum 256bit integers and fixed point decimals. It works correctly on Postgres and BigQuery. All the other destinations have insufficient precision.

complex data type tells dlt to load that element as JSON or struct and do not attempt to flatten or create a child table out of it.

time data type is saved in destination without timezone info, if timezone is included it is stripped. E.g. '14:01:02+02:00 -> '14:01:02'.

tip

The precision and scale are interpreted by particular destination and are validated when a column is created. Destinations that do not support precision for a given data type will ignore it.

The precision for timestamp is useful when creating parquet files. Use 3 - for milliseconds, 6 for microseconds, 9 for nanoseconds

The precision for bigint is mapped to available integer types ie. TINYINT, INT, BIGINT. The default is 64 bits (8 bytes) precision (BIGINT)

Schema settings

The settings section of schema file lets you define various global rules that impact how tables and columns are inferred from data.

💡 It is the best practice to use those instead of providing the exact column schemas via columns argument or by pasting them in yaml.

Data type autodetectors

You can define a set of functions that will be used to infer the data type of the column from a value. The functions are run from top to bottom on the lists. Look in detections.py to see what is available.

settings:
  detections:
    - timestamp
    - iso_timestamp
    - iso_date
    - large_integer
    - hexbytes_to_text
    - wei_to_double

Column hint rules

You can define a global rules that will apply hints of a newly inferred columns. Those rules apply to normalized column names. You can use column names directly or with regular expressions.

Example from ethereum schema:

settings:
  default_hints:
    foreign_key:
      - _dlt_parent_id
    not_null:
      - re:^_dlt_id$
      - _dlt_root_id
      - _dlt_parent_id
      - _dlt_list_idx
    unique:
      - _dlt_id
    cluster:
      - block_hash
    partition:
      - block_timestamp

Preferred data types

You can define rules that will set the data type for newly created columns. Put the rules under preferred_types key of settings. On the left side there's a rule on a column name, on the right side is the data type.

❗See the column hint rules for naming convention!

Example:

settings:
  preferred_types:
    timestamp: timestamp
    re:^inserted_at$: timestamp
    re:^created_at$: timestamp
    re:^updated_at$: timestamp
    re:^_dlt_list_idx$: bigint

Applying data types directly with @dlt.resource and apply_hints

dlt offers the flexibility to directly apply data types and hints in your code, bypassing the need for importing and adjusting schemas. This approach is ideal for rapid prototyping and handling data sources with dynamic schema requirements.

Direct specification in @dlt.resource

Directly define data types and their properties, such as nullability, within the @dlt.resource decorator. This eliminates the dependency on external schema files. For example:

@dlt.resource(name='my_table', columns={"my_column": {"data_type": "bool", "nullable": True}})
def my_resource():
    for i in range(10):
        yield {'my_column': i % 2 == 0}

This code snippet sets up a nullable boolean column named my_column directly in the decorator.

Using apply_hints

When dealing with dynamically generated resources or needing to programmatically set hints, apply_hints is your tool. It's especially useful for applying hints across various collections or tables at once.

For example, to apply a complex data type across all collections from a MongoDB source:

all_collections = ["collection1", "collection2", "collection3"]  # replace with your actual collection names
source_data = mongodb().with_resources(*all_collections)

for col in all_collections:
    source_data.resources[col].apply_hints(columns={"column_name": {"data_type": "complex"}})

pipeline = dlt.pipeline(
    pipeline_name="mongodb_pipeline",
    destination="duckdb",
    dataset_name="mongodb_data"
)
load_info = pipeline.run(source_data)

This example iterates through MongoDB collections, applying the complex data type to a specified column, and then processes the data with pipeline.run.

Export and import schema files

Please follow the guide on how to adjust a schema to export and import yaml schema files in your pipeline.

Attaching schemas to sources

We recommend to not create schemas explicitly. Instead, user should provide a few global schema settings and then let the table and column schemas to be generated from the resource hints and the data itself.

The dlt.source decorator accepts a schema instance that you can create yourself and modify in whatever way you wish. The decorator also support a few typical use cases:

Schema created implicitly by decorator

If no schema instance is passed, the decorator creates a schema with the name set to source name and all the settings to default.

Automatically load schema file stored with source python module

If no schema instance is passed, and a file with a name {source name}_schema.yml exists in the same folder as the module with the decorated function, it will be automatically loaded and used as the schema.

This should make easier to bundle a fully specified (or pre-configured) schema with a source.

Schema is modified in the source function body

What if you can configure your schema or add some tables only inside your schema function, when i.e. you have the source credentials and user settings available? You could for example add detailed schemas of all the database tables when someone requests a table data to be loaded. This information is available only at the moment source function is called.

Similarly to the source_state() and resource_state() , source and resource function has current schema available via dlt.current.source_schema().

Example:

@dlt.source
def textual(nesting_level: int):
    # get the source schema from the `current` context
    schema = dlt.current.source_schema()
    # remove date detector
    schema.remove_type_detection("iso_timestamp")
    # convert UNIX timestamp (float, withing a year from NOW) into timestamp
    schema.add_type_detection("timestamp")
    schema._compile_settings()

    return dlt.resource([])