Configuration Keys¶

This page provides a consolidated reference of every configuration field across all weevr models. Fields are grouped by the model they belong to.

For YAML syntax and complete examples, see the schema reference pages: Thread | Weave | Loom

Thread¶

The top-level configuration unit for a single data pipeline.

Field	Type	Default	Description
`config_version`	`str`	required	Schema version identifier
`name`	`str`	`""`	Thread name
`sources`	`dict[str, Source]`	required	Named data sources
`steps`	`list[Step]`	`[]`	Ordered transformation pipeline
`target`	`Target`	required	Output destination
`write`	`WriteConfig`	`None`	Write mode and merge settings
`keys`	`KeyConfig`	`None`	Key management configuration
`validations`	`list[ValidationRule]`	`None`	Pre-write data quality rules
`assertions`	`list[Assertion]`	`None`	Post-execution assertions
`load`	`LoadConfig`	`None`	Incremental load settings
`tags`	`list[str]`	`None`	Free-form tags
`params`	`dict[str, ParamSpec]`	`None`	Parameter declarations
`defaults`	`dict[str, Any]`	`None`	Inherited defaults
`failure`	`FailureConfig`	`None`	Failure handling policy
`execution`	`ExecutionConfig`	`None`	Runtime settings
`cache`	`bool`	`None`	Cache DataFrame before writing
`exports`	`list[Export]`	`None`	Secondary output destinations
`lookups`	`dict[str, Lookup]`	`None`	Thread-level lookup definitions
`column_sets`	`dict[str, ColumnSet]`	`None`	Thread-level column set definitions
`variables`	`dict[str, VariableSpec]`	`None`	Thread-level variable declarations
`pre_steps`	`list[HookStep]`	`None`	Hook steps run before thread execution
`post_steps`	`list[HookStep]`	`None`	Hook steps run after thread execution
`audit_templates`	`dict[str, dict[str, str]]`	`None`	Named audit column templates available to all targets
`connections`	`dict[str, OneLakeConnection]`	`None`	Named connection definitions
`with`	`dict[str, SubPipeline]`	`None`	Named sub-pipelines (CTEs) resolved before main steps

Export¶

A named secondary output destination for thread data. Exports write the same post-mapping, audit-injected DataFrame as the primary target.

Field	Type	Default	Description
`name`	`str`	required	Unique identifier (valid Python identifier)
`description`	`str`	`None`	Human-readable label shown in `explain()`
`type`	`str`	required	Output format: `delta`, `parquet`, `csv`, `json`, `orc`
`path`	`str`	`None`	OneLake path. Required when `alias` is not set. Supports context variables.
`alias`	`str`	`None`	Metastore alias. Delta type only, mutually exclusive with `path`.
`mode`	`str`	`"overwrite"`	Write mode (only `overwrite` in v1)
`partition_by`	`list[str]`	`None`	Partition columns (independent of primary target)
`on_failure`	`str`	`"warn"`	`abort` fails the thread; `warn` logs and continues
`enabled`	`bool`	`True`	Set to `false` to suppress an inherited export
`options`	`dict[str, str]`	`None`	Format-specific Spark DataFrameWriter options

Related: Exports guide

AuditTemplate¶

A named set of audit columns defined under audit_templates at loom, weave, or thread level. Each template is a dict[str, str] mapping column names to Spark SQL expressions.

audit_templates:
  lineage:
    _source_system: "${param.source_system}"
    _load_ts: "current_timestamp()"
    _run_id: "${param.run_id}"

Templates are referenced by name in target.audit_template. The built-in presets fabric and minimal do not require a declaration.

Related: Audit Templates guide

Source¶

A data source declaration referenced by alias within a thread. A source is a direct data reference (with type), a lookup reference (with lookup), a connection-based reference (with connection + table), or a generated sequence (with type set to date_sequence or int_sequence).

Field	Type	Default	Description
`type`	`str`	conditional	Source type: `delta`, `csv`, `json`, `parquet`, `excel`, `date_sequence`, `int_sequence`. Required when neither `lookup` nor `connection` is set.
`lookup`	`str`	`None`	Lookup name (thread-, weave-, or loom-level). Mutually exclusive with `type`.
`alias`	`str`	`None`	Table alias (required for `delta` without `connection`). Mutually exclusive with `connection`.
`path`	`str`	`None`	File path (required for file-based types)
`options`	`dict[str, Any]`	`{}`	Spark reader options
`dedup`	`DedupConfig`	`None`	Post-read deduplication
`connection`	`str`	`None`	Named connection reference. Requires `table`.
`schema`	`str`	`None`	Schema override within the connection's lakehouse
`table`	`str`	`None`	Table name (required when `connection` is set)
`start`	`str \\| int`	`None`	Range start, inclusive (required for generated types)
`end`	`str \\| int`	`None`	Range end, inclusive (required for generated types)
`column`	`str`	`None`	Output column name (required for generated types)
`step`	`str \\| int`	`None`	Step interval: `day`/`week`/`month`/`year` for dates, positive int for `int_sequence`

DedupConfig¶

Field	Type	Default	Description
`keys`	`list[str]`	required	Deduplication grouping columns
`order_by`	`str`	`None`	Sort expression for row selection

Target¶

Write destination with column mapping and partitioning.

Field	Type	Default	Description
`alias`	`str`	`None`	Lakehouse table alias
`path`	`str`	`None`	File path for file-based targets
`connection`	`str`	`None`	Named connection reference. Requires `table`. Mutually exclusive with `alias`.
`schema`	`str`	`None`	Schema override within the connection's lakehouse
`table`	`str`	`None`	Table name within the connection's lakehouse. Requires `connection`.
`mapping_mode`	`"auto" \\| "explicit"`	`"auto"`	Column mapping strategy
`columns`	`dict[str, ColumnMapping]`	`None`	Per-column mapping
`partition_by`	`list[str]`	`None`	Partition columns
`audit_columns`	`dict[str, str]`	`None`	Audit column definitions as name-expression pairs
`audit_template`	`str \\| list[str]`	`None`	Template name(s) to apply. Accepts built-in presets (`fabric`, `minimal`) or user-defined names.
`audit_template_inherit`	`bool`	`True`	When `False`, suppresses inherited `audit_template` from parent levels
`audit_columns_exclude`	`list[str]`	`None`	Column names or glob patterns excluded from the resolved template set
`naming`	`NamingConfig`	`None`	Naming normalization
`warp`	`str \\| false \\| null`	`null`	Warp reference, opt-out, or auto-discover
`warp_mode`	`"auto" \\| null`	`null`	Auto-generate warp after write
`warp_init`	`bool`	`false`	Pre-initialize table from warp
`warp_enforcement`	`"warn" \\| "enforce" \\| "off"`	`"warn"`	Warp contract enforcement mode
`schema_drift`	`"lenient" \\| "strict" \\| "adaptive"`	`"lenient"`	Schema drift handling mode
`on_drift`	`"error" \\| "warn" \\| "ignore"`	`"warn"`	Severity for strict drift mode
`dimension`	`DimensionConfig`	`None`	SCD Type 2 behavior. Mutually exclusive with `fact`. See thread schema: target.dimension.
`fact`	`FactConfig`	`None`	Fact table FK validation. Mutually exclusive with `dimension`. See thread schema: target.fact.
`seed`	`SeedConfig`	`None`	Static seed rows inserted on initial load. See thread schema: target.seed.

ColumnMapping¶

Field	Type	Default	Description
`expr`	`SparkExpr`	`None`	Expression for column value (mutually exclusive with `drop`)
`type`	`str`	`None`	Target data type for casting
`default`	`Any`	`None`	Default when source is null
`drop`	`bool`	`False`	Drop column from output (mutually exclusive with `expr`)

WriteConfig¶

Write mode and merge behavior for the target.

Field	Type	Default	Description
`mode`	`"overwrite" \\| "append" \\| "merge"`	`"overwrite"`	Write strategy
`match_keys`	`list[str]`	`None`	Merge match keys (required for `merge` mode)
`on_match`	`"update" \\| "ignore"`	`"update"`	Action when source row matches target
`on_no_match_target`	`"insert" \\| "ignore"`	`"insert"`	Action for new source rows
`on_no_match_source`	`"delete" \\| "soft_delete" \\| "ignore"`	`"ignore"`	Action for missing source rows
`soft_delete_column`	`str`	`None`	Column for soft delete flag (required for `soft_delete`)
`soft_delete_value`	`bool`	`True`	Value written to the soft delete column when a row is soft-deleted
`soft_delete_active_value`	`bool`	`None`	Value written to the soft delete column for active (non-deleted) rows

LoadConfig¶

Incremental load mode and watermark tracking.

Field	Type	Default	Description
`mode`	`"full" \\| "incremental_watermark" \\| "incremental_parameter" \\| "cdc"`	`"full"`	Load strategy
`watermark_column`	`str`	`None`	Column for watermark comparison. Required for `incremental_watermark`; optional for `cdc` mode when `cdc.operation_column` is set (rejected when `cdc.preset` is `"delta_cdf"`).
`watermark_type`	`"timestamp" \\| "date" \\| "int" \\| "long"`	`None`	Watermark column data type. Required whenever `watermark_column` is set.
`watermark_inclusive`	`bool`	`False`	Include rows equal to last watermark
`watermark_store`	`WatermarkStoreConfig`	`None`	Watermark persistence backend
`cdc`	`CdcConfig`	`None`	CDC configuration
`watermark_format`	`str`	`None`	Spark `DateTimeFormatter` pattern used to parse a string-typed `watermark_column`. Only valid with `watermark_type` of `"timestamp"` or `"date"`.

Cross-field rules enforced at parse time:

mode == "incremental_watermark" requires watermark_column
mode == "cdc" requires cdc config
mode == "cdc" with cdc.preset == "delta_cdf" rejects watermark_column (CDF uses commit-version tracking automatically)
mode == "cdc" with cdc.operation_column (generic CDC) may compose with watermark_column to narrow the read window for append-only CDC history tables; watermark_type is required whenever watermark_column is set in cdc mode
watermark_format requires watermark_column and requires watermark_type to be "timestamp" or "date"

WatermarkStoreConfig¶

Field	Type	Default	Description
`type`	`"table_properties" \\| "metadata_table"`	`"table_properties"`	Storage backend
`table_path`	`str`	`None`	Path for metadata table (required for `metadata_table`)

CdcConfig¶

Field	Type	Default	Description
`preset`	`"delta_cdf"`	`None`	CDC preset (mutually exclusive with `operation_column`)
`operation_column`	`str`	`None`	Operation type column (mutually exclusive with `preset`)
`insert_value`	`str`	`None`	Value indicating insert
`update_value`	`str`	`None`	Value indicating update
`delete_value`	`str`	`None`	Value indicating delete
`on_delete`	`"hard_delete" \\| "soft_delete"`	`"hard_delete"`	Delete handling strategy

KeyConfig¶

Business key, surrogate key, and change detection settings.

Field	Type	Default	Description
`business_key`	`list[str]`	`None`	Natural business key columns
`surrogate_key`	`SurrogateKeyConfig`	`None`	Surrogate key generation
`change_detection`	`ChangeDetectionConfig`	`None`	Change detection hash

SurrogateKeyConfig¶

Field	Type	Default	Description
`name`	`str`	required	Output column name
`algorithm`	`"xxhash64" \\| "sha1" \\| "sha256" \\| "sha384" \\| "sha512" \\| "md5" \\| "crc32" \\| "murmur3"`	`"sha256"`	Hash algorithm
`output`	`"native" \\| "string"`	`"native"`	Output type for integer-returning algorithms (xxhash64, crc32, murmur3). `native` preserves the return type; `string` casts to StringType.

ChangeDetectionConfig¶

Field	Type	Default	Description
`name`	`str`	required	Output column name
`columns`	`list[str]`	required	Columns included in the hash
`algorithm`	`"xxhash64" \\| "sha1" \\| "sha256" \\| "sha384" \\| "sha512" \\| "md5" \\| "crc32" \\| "murmur3"`	`"md5"`	Hash algorithm
`output`	`"native" \\| "string"`	`"native"`	Output type for integer-returning algorithms. `native` preserves the return type; `string` casts to StringType.

ValidationRule¶

Pre-write data quality rules.

Field	Type	Default	Description
`name`	`str`	required	Rule name
`rule`	`SparkExpr`	required	Spark SQL boolean expression
`severity`	`"info" \\| "warn" \\| "error" \\| "fatal"`	`"error"`	Failure severity

Assertion¶

Post-execution assertions on the target dataset.

Field	Type	Default	Description
`type`	`"row_count" \\| "column_not_null" \\| "unique" \\| "expression" \\| "fk_sentinel_rate"`	required	Assertion type
`severity`	`"info" \\| "warn" \\| "error" \\| "fatal"`	`"warn"`	Failure severity
`columns`	`list[str]`	`None`	Columns for `column_not_null` and `unique`
`column`	`str`	`None`	FK column for `fk_sentinel_rate`
`min`	`int`	`None`	Minimum for `row_count`
`max`	`int`	`None`	Maximum for `row_count`
`expression`	`SparkExpr`	`None`	Expression for `expression` type
`sentinel`	`Any`	`None`	Single sentinel value for `fk_sentinel_rate`
`sentinels`	`list[Any]`	`None`	Multiple sentinel values for `fk_sentinel_rate`
`max_rate`	`float`	`None`	Maximum acceptable sentinel rate (0.0-1.0) for `fk_sentinel_rate`
`message`	`str`	`None`	Custom failure message for `fk_sentinel_rate`

ExecutionConfig¶

Runtime settings that cascade through loom/weave/thread.

Field	Type	Default	Description
`log_level`	`"minimal" \\| "standard" \\| "verbose" \\| "debug"`	`"standard"`	Logging verbosity
`trace`	`bool`	`True`	Collect execution spans

FailureConfig¶

Per-thread failure handling policy.

Field	Type	Default	Description
`on_failure`	`"abort_weave" \\| "skip_downstream" \\| "continue"`	`"abort_weave"`	Failure policy

ParamSpec¶

Typed parameter declaration.

Field	Type	Default	Description
`name`	`str`	required	Parameter name
`type`	`str`	required	Data type: `string`, `int`, `float`, `bool`, `date`, `timestamp`, `list[string]`
`required`	`bool`	`True`	Whether the parameter must be supplied
`default`	`Any`	`None`	Default value
`description`	`str`	`""`	Human-readable description

NamingConfig¶

Column and table naming normalization. Cascades through loom/weave/thread/target.

Field	Type	Default	Description
`columns`	`NamingPattern`	`None`	Column naming pattern
`tables`	`NamingPattern`	`None`	Table naming pattern
`exclude`	`list[str]`	`[]`	Names or patterns excluded from normalization
`on_collision`	`Literal["suffix", "error"]`	`"error"`	Duplicate name handling after normalization
`reserved_words`	`ReservedWordConfig`	`None`	Reserved word protection config

Supported NamingPattern values: snake_case, camelCase, PascalCase, UPPER_SNAKE_CASE, Title_Snake_Case, Title Case, lowercase, UPPERCASE, kebab-case, none.

ReservedWordConfig¶

Field	Type	Default	Description
`strategy`	`Literal[7 values]`	`"quote"`	How to handle reserved words
`prefix`	`str`	`"_"`	Prefix when strategy is `prefix`
`suffix`	`str`	`"_col"`	Suffix when strategy is `suffix`
`rename_map`	`dict[str, str]`	`None`	Explicit collision map for `rename`
`fallback`	`Literal[6 values]`	`"quote"`	Fallback for unmapped `rename` collisions
`preset`	`str \\| list`	`None`	Built-in word list presets
`extend`	`list[str]`	`[]`	Extra words to treat as reserved
`exclude`	`list[str]`	`[]`	Words to remove from check

Strategies¶

Strategy	Behavior
`quote`	Keep name as-is; Spark backtick-quotes in SQL
`prefix`	Prepend `prefix` to colliding names
`suffix`	Append `suffix` to colliding names
`error`	Raise `ConfigError` listing all collisions
`rename`	Apply `rename_map`; unmapped use `fallback`
`revert`	Discard the rename, keep pre-normalization name
`drop`	Remove colliding columns (not valid for tables)

The drop strategy is not valid for table names — neither as a direct strategy nor as a fallback. Both raise ConfigError at runtime in normalize_table_name.

Presets¶

The preset field selects one or more built-in reserved word lists. When omitted, the ANSI SQL list is used as the default. Specifying any preset replaces that default — to include ANSI words alongside other presets, list ansi explicitly.

A single string is accepted as shorthand for a one-element list.

Preset	Description
`ansi`	ANSI SQL reserved keywords (~80 words)
`dax`	DAX reserved words for Power BI semantic models
`m`	M language (Power Query) reserved words
`powerbi`	Convenience alias — expands to `dax` + `m`
`tsql`	T-SQL reserved keywords for SQL endpoints

Each preset is self-contained: preset: [tsql] alone gives full T-SQL protection without needing to add ansi.

# Default (no preset = ANSI):
reserved_words:
  strategy: prefix

# Suffix — append string to collisions:
reserved_words:
  strategy: suffix
  suffix: "_col"       # order -> order_col

# Explicit rename map with fallback:
reserved_words:
  strategy: rename
  rename_map:
    order: order_name
    select: select_col
  fallback: error      # unmapped collisions fail

# Revert — keep original name on collision:
reserved_words:
  strategy: revert

# Drop — remove colliding columns:
reserved_words:
  strategy: drop

# Power BI convenience alias (= dax + m):
reserved_words:
  preset: powerbi
  strategy: prefix

# Multi-engine (ANSI + DAX + T-SQL):
reserved_words:
  preset: [ansi, dax, tsql]
  strategy: prefix
  extend: [fiscal_year]
  exclude: [by]

extend and exclude compose on top of the resolved preset union, adding or removing individual words.

Weave¶

A collection of threads with shared lookups, hooks, and defaults.

Field	Type	Default	Description
`config_version`	`str`	required	Schema version identifier
`name`	`str`	`""`	Weave name
`threads`	`list[ThreadEntry]`	required	Thread references
`lookups`	`dict[str, Lookup]`	`None`	Named lookup sources shared across threads
`variables`	`dict[str, VariableSpec]`	`None`	Weave-scoped typed variables
`pre_steps`	`list[HookStep]`	`None`	Hook steps run before threads
`post_steps`	`list[HookStep]`	`None`	Hook steps run after threads
`defaults`	`dict[str, Any]`	`None`	Default values cascaded to threads
`params`	`dict[str, ParamSpec]`	`None`	Parameter declarations
`execution`	`ExecutionConfig`	`None`	Runtime settings cascaded to threads
`naming`	`NamingConfig`	`None`	Naming normalization cascaded to threads
`column_sets`	`dict[str, ColumnSet]`	`None`	Named column sets for bulk rename
`audit_templates`	`dict[str, dict[str, str]]`	`None`	Named audit column templates cascaded to threads
`connections`	`dict[str, OneLakeConnection]`	`None`	Named connection definitions cascaded to threads

ThreadEntry¶

Field	Type	Default	Description
`name`	`str`	`""`	Thread name (required for inline definitions)
`ref`	`str`	`None`	Path to an external `.thread` file
`dependencies`	`list[str]`	`None`	Explicit upstream thread names
`condition`	`ConditionSpec`	`None`	Conditional execution gate

ConditionSpec¶

Field	Type	Default	Description
`when`	`str`	required	Condition expression. Supports `${param.x}` references, `table_exists()`, `table_empty()`, `row_count()`, and boolean operators.

Loom¶

Deployment unit grouping one or more weaves.

Field	Type	Default	Description
`config_version`	`str`	required	Schema version identifier
`name`	`str`	`""`	Loom name
`weaves`	`list[WeaveEntry]`	required	Weave references
`defaults`	`dict[str, Any]`	`None`	Default values cascaded to weaves and threads
`params`	`dict[str, ParamSpec]`	`None`	Parameter declarations
`execution`	`ExecutionConfig`	`None`	Runtime settings cascaded down
`naming`	`NamingConfig`	`None`	Naming normalization cascaded down
`column_sets`	`dict[str, ColumnSet]`	`None`	Named column sets cascaded to weaves
`lookups`	`dict[str, Lookup]`	`None`	Loom-level lookup definitions cascaded to weaves
`variables`	`dict[str, VariableSpec]`	`None`	Loom-level variable declarations
`pre_steps`	`list[HookStep]`	`None`	Hook steps run before any weave executes
`post_steps`	`list[HookStep]`	`None`	Hook steps run after all weaves complete
`audit_templates`	`dict[str, dict[str, str]]`	`None`	Named audit column templates cascaded to weaves and threads
`connections`	`dict[str, OneLakeConnection]`	`None`	Named connection definitions cascaded to weaves and threads

WeaveEntry¶

Field	Type	Default	Description
`name`	`str`	`""`	Weave name (required for inline definitions)
`ref`	`str`	`None`	Path to an external `.weave` file
`condition`	`ConditionSpec`	`None`	Conditional execution gate

OneLakeConnection¶

Named connection declaration for Fabric OneLake sources and targets. Declared in the connections: block at loom, weave, or thread level and cascaded through the standard merge rules.

Field	Type	Default	Description
`type`	`Literal["onelake"]`	required	Connection type identifier. Only `onelake` is supported.
`workspace`	`str`	required	OneLake workspace GUID or `${param.*}` variable reference
`lakehouse`	`str`	required	OneLake lakehouse GUID or `${param.*}` variable reference
`default_schema`	`str`	`None`	Default schema applied to tables in this connection when none is set

workspace and lakehouse both support ${param.*} and ${variable.*} substitution so one declaration can serve multiple environments.

ColumnSet¶

Named column mapping sourced from Delta, YAML, or runtime parameters.

Field	Type	Default	Description
`source`	`ColumnSetSource`	`None`	External source definition
`param`	`str`	`None`	Runtime parameter name (mutually exclusive with `source`)
`on_unmapped`	`Literal["pass_through", "error"]`	`"pass_through"`	Behavior for unmapped DataFrame columns
`on_extra`	`Literal["ignore", "warn", "error"]`	`"ignore"`	Behavior for extra mapping keys
`on_failure`	`Literal["abort", "warn", "skip"]`	`"abort"`	Behavior on source read failure

ColumnSetSource¶

A delta source can be addressed two ways: by registered table alias (default Spark catalog) or by named connection plus table (resolves through the lakehouse referenced by a connections: entry on the loom or weave). The two modes are mutually exclusive.

Field	Type	Default	Description
`type`	`Literal["delta", "yaml"]`	required	Source type
`alias`	`str`	`None`	Registered table alias. One of `alias` or `connection` is required when `type="delta"`. Mutually exclusive with `connection`.
`path`	`str`	`None`	File path to the YAML mapping file. Required when `type="yaml"`.
`connection`	`str`	`None`	Reference to a named connection defined at the loom or weave level. Only valid for `type="delta"`. Mutually exclusive with `alias`.
`schema`	`str`	`None`	Schema name within the connection's lakehouse. Only valid alongside `connection`.
`table`	`str`	`None`	Table name within the connection's lakehouse. Required when `connection` is set.
`from_column`	`str`	`"source_name"`	Column containing raw names
`to_column`	`str`	`"target_name"`	Column containing target names
`filter`	`str`	`None`	SQL WHERE expression applied before collecting mappings

Lookup¶

A weave-level named data definition referenced by threads via source.lookup.

Field	Type	Default	Description
`source`	`Source`	required	Source definition for the lookup data
`materialize`	`bool`	`False`	Pre-read and cache/broadcast before thread execution
`strategy`	`"broadcast" \\| "cache"`	`"cache"`	Materialization strategy
`key`	`list[str]`	`None`	Join key columns for narrow projection
`values`	`list[str]`	`None`	Payload columns to retain (requires `key`)
`filter`	`str`	`None`	SQL WHERE expression applied before projection
`unique_key`	`bool`	`False`	Validate key uniqueness after filtering
`on_failure`	`"abort" \\| "warn"`	`"abort"`	Behavior on duplicate keys (only valid when `unique_key` is true)

HookStep¶

Discriminated union of hook step types, dispatched on the type field. Used in weave pre_steps and post_steps.

QualityGateStep¶

Field	Type	Default	Description
`type`	`"quality_gate"`	required	Step type discriminator
`name`	`str`	`None`	Step name for telemetry
`on_failure`	`"abort" \\| "warn"`	phase default	Failure behavior
`check`	`str`	required	Check type: `source_freshness`, `row_count_delta`, `row_count`, `table_exists`, `expression`
`source`	`str`	`None`	Table alias (for `source_freshness`, `table_exists`)
`max_age`	`str`	`None`	Duration string (for `source_freshness`)
`target`	`str`	`None`	Table alias (for `row_count_delta`, `row_count`)
`max_decrease_pct`	`float`	`None`	Max decrease % (for `row_count_delta`)
`max_increase_pct`	`float`	`None`	Max increase % (for `row_count_delta`)
`min_delta`	`int`	`None`	Min absolute row change (for `row_count_delta`)
`max_delta`	`int`	`None`	Max absolute row change (for `row_count_delta`)
`min_count`	`int`	`None`	Min row count (for `row_count`)
`max_count`	`int`	`None`	Max row count (for `row_count`)
`sql`	`str`	`None`	SQL expression (for `expression`)
`message`	`str`	`None`	Failure message (for `expression`)

SqlStatementStep¶

Field	Type	Default	Description
`type`	`"sql_statement"`	required	Step type discriminator
`name`	`str`	`None`	Step name for telemetry
`on_failure`	`"abort" \\| "warn"`	phase default	Failure behavior
`sql`	`str`	required	Spark SQL statement to execute
`set_var`	`str`	`None`	Variable name to capture scalar result into

LogMessageStep¶

Field	Type	Default	Description
`type`	`"log_message"`	required	Step type discriminator
`name`	`str`	`None`	Step name for telemetry
`on_failure`	`"abort" \\| "warn"`	phase default	Failure behavior
`message`	`str`	required	Message template (supports `${var.name}`)
`level`	`"info" \\| "warn" \\| "error"`	`"info"`	Log level

VariableSpec¶

Weave-scoped variable declaration. Set by hook steps via set_var, referenced as ${var.name} in config expressions.

Field	Type	Default	Description
`type`	`str`	required	Scalar type: `string`, `int`, `long`, `float`, `double`, `boolean`, `timestamp`, `date`
`default`	`str \\| int \\| float \\| bool`	`None`	Default value when no hook sets the variable