dpmxl.columnar¶

Return the canonical text of an item: category:code or code.

Returns:

• str ( str ) –

  The canonical form used in txt and open key columns.

Examples:

>>> from dpmxl import Item
>>> canonical_item(Item.parse("eba_qEC:qx01"))
'qEC:qx01'

Split a scalar value into the four value columns.

Numbers fill num and thr; Booleans flag; text, items (canonical, see canonical_item) and dates (ISO) txt. None leaves all four null.

Returns:

• dict[str, Any] –

  dict[str, Any]: num, thr, txt and flag entries.

Examples:

>>> from dpmxl import Number
>>> value_columns(Number.of("1.5", 1))
{'num': Decimal('1.5'), 'thr': Decimal('0.05'), 'txt': None, 'flag': None}
>>> value_columns(True)
{'num': None, 'thr': None, 'txt': None, 'flag': True}

Raises:

• ColumnarUnsupported –

  For a number or threshold that Decimal(38, 12) does not hold exactly.

Build a key-less one-row frame holding one value.

Returns:

• Frame ( Frame ) –

  The scalar, with the columns of

• Frame –

  value_columns.

Adapt data points to the frame layout of a FrameProvider.

Builds the columns described by FrameProvider from DataPoint objects; cell holds repr(point.ref).

Parameters:

• points (Sequence[DataPoint]) –

  Selected data points.

• open_keys (Sequence[str], default: () ) –

  The table's open key names, kept even without points.

Returns:

• DataFrame –

  pl.DataFrame: One row per point.

Raises:

• ColumnarUnsupported –

  For numeric or Boolean key values, items without a category, or numbers Decimal(38, 12) does not hold.

Vectorized normalize_code.

Returns:

• Expr –

  pl.Expr: Numeric codes zero-padded to four digits, others upper-case.

Evaluate a validation rule on frames, honouring its precondition.

The columnar counterpart of evaluate_rule: the same statuses and counts, and failing keys in canonical form (see canonical_key). Syntax and evaluation errors are returned as status error, not raised. It may run concurrently from several threads with a shared cache.

Parameters:

• expression (str | Node) –

  Rule expression text or a parsed tree.

• provider (FrameProvider) –

  Columnar report data.

• precondition (str | Node | None, default: None ) –

  Optional precondition; the rule is not applicable unless it yields at least one true record.

• cache (FrameCache | None, default: None ) –

  Selections shared by the rules of one report; a new cache per call when omitted.

• max_failing_keys (int, default: 1000 ) –

  How many failing keys to list in failing_keys; failing_frame has them all.

Returns:

• ColumnarResult ( ColumnarResult ) –

  Status, record counts and the keys of failing records.

Raises:

• ColumnarUnsupported –

  When the reference evaluator must run instead.

Examples:

Fall back to the reference evaluator when needed:

>>> from dpmxl import MemoryProvider, Number, TableLayout, evaluate_rule
>>> provider = MemoryProvider(
...     [TableLayout("T", rows=("0010",), columns=("0010",))],
...     {("T", "0010", "0010", None, ()): Number.of(3, 0)},
... )
>>> def run(expression):
...     try:
...         return evaluate_rule_columnar(
...             expression, PointsFrameProvider(provider)
...         ).status
...     except ColumnarUnsupported:
...         return evaluate_rule(expression, provider).status
>>> run("{tT, r0010, c0010} > 0")
'passed'
>>> run("year(#2026-12-31#) = 2026")  # dates: reference evaluator
'passed'

Canonicalize a reference outcome key for comparison with columnar keys.

Closed coordinates are normalized with normalize_code; other key values become the text the columnar evaluator uses (items as category:code).

Returns:

• dict[str, str | None] –

  dict[str, str | None]: Key name to canonical string.

Examples:

>>> canonical_key({"r": "10", "qCPI": "LEI-A"})
{'r': '0010', 'qCPI': 'LEI-A'}

Return whether a cell's coordinates are selected (for providers).

Returns:

• bool ( bool ) –

  Whether every axis matches.