GAMMA.INV silently returned the first bisection midpoint for alpha in
~[143, 171.62] (e.g. GAMMA.INV(0.5, 143, 1) gave 358.0 instead of
142.667) and #NUM! above that, because the Newton-step pdf evaluates
Gamma(a), b**a and value**(a-1) in linear domain, all of which overflow
even though the density itself is a small representable number. The same
pattern breaks the GAMMA.DIST, CHISQ.DIST and F.DIST densities and
GAMMALN, which computed log(Gamma(x)) through Gamma(x).
Evaluate these in log domain via the existing logGamma, and scale the
incomplete-gamma series/continued-fraction iteration cap as O(sqrt(a)),
which both expansions need to converge near x ~ a once the shape is
large (GAMMA.INV drifted from the true quantile above alpha ~5000 and
returned alpha+1 by alpha=10000; same for CHISQ.INV at high df).
calculateInverse() treated a CDF value of exactly 0.0 as "the guess is the
root" and collapsed the bracket with $b = $a, ending the search. inverse()
already rejects probability <= 0, so a CDF of 0.0 inside the bisection can
only be a float64 underflow at a guess far below the root - never an exact
hit. It is just an ordinary "guess too low" and belongs in the existing
else branch.
The underflow is reached whenever a probe lands many standard deviations
from the mean, so the search collapses after two or three iterations and
returns whichever midpoint it was holding. BETAINV(0.5, 5000, 5000) gave
0.25, though Beta(a, a) is symmetric and its median is exactly 0.5;
BETAINV(0.5, 20000, 3) gave 0.5 against a true 0.99986. Onset is around
alpha = 1080 for beta = 1, where the closed form 0.5 ** (1 / alpha) is
available to check against.
That test also happened to stop the search for shapes where incompleteBeta
declines to evaluate at all and returns 0 for every x, so inverse() now
rejects alpha + beta above that documented limit up front: the CDF is
identically zero there, so no quantile exists to search for.
Tests cover a shape x probability grid against scipy reference values plus
three checks that need no external oracle: the symmetric median, the
Beta(alpha, 1) closed form, and the BETADIST round trip / mirror identity.
Merging #4946 (bracket expansion) with #4945 (accurate incomplete
gamma) means p=0.9999999 no longer hits the alpha*beta*5 safety
ceiling -- the corrected regularizedGammaP/Q resolve the true root.
Fold it into the extreme-tail data provider against the closed-form
-ln(1-p) check instead of asserting the stale clamped bound.
It has no remaining callers (kept only for BC) so nothing else exercises
it. Values checked against mpmath's gammainc(a, 0, x); the deep-tail case
(a=4, x=80) reproduces the convergence bug this PR fixes.
The incomplete gamma primitive used a fixed 32-term power series with no
convergence test, so GAMMA.DIST, GAMMADIST, CHISQ.DIST(.RT), GAMMAINV and
CHISQ.INV were grossly wrong once the series argument reached ~32
(e.g. CHISQ.DIST.RT(80, 4) returned 0.806 instead of 1.74e-16).
Replace it with the standard convergence-tested regularized incomplete
gamma: series P(a,x) for x < a+1, continued fraction Q(a,x) for x >= a+1.
CHISQ.DIST.RT now uses Q directly so the right tail stays free of
1 - P cancellation. Consolidates the duplicate copy that already existed
privately in ChiSquared onto the shared primitive.
Excel reported sample 3's ByQuarter sheet as corrupt: a date field
group (rangePr groupBy="quarters") was emitted without the date bounds
Excel requires. When no explicit start/end date is supplied, fall back
to the sentinel range (1900-01-01 .. 9999-12-31) so both the sharedItems
(minDate/maxDate) and the rangePr (startDate/endDate) are always present,
matching the <1/1/1900 / >12/31/9999 out-of-range group items.
Also fixes ordered_imports in Worksheet.php (php-cs-fixer CI failure):
the PivotTable\PivotTable use statement is moved before the Sparkline
imports.
Cover the previously-unexercised code paths flagged by Coveralls:
- writer: month and year date grouping, and a fractional numeric interval
(the non-integer num() path);
- reader: reading an explicit dataField subtotal attribute back from a saved
file;
- worksheet: the getPivotTables() alias, addPivotTable(), case-insensitive
getPivotTableByName(), the no-match lookup, and removePivotTableCollection().
Cover the PivotFieldGroup, PivotField, PivotCacheDefinition and PivotTable
accessors that were not exercised by the reader/builder tests (numeric and
date grouping getters, field-group storage on the cache definition, pivot
field setters, and PivotTable::__toString), restoring code coverage.
Use getSheetByNameOrThrow() instead of the nullable getSheetByName() for
chained calls, assert non-null before dereferencing getCacheDefinition() and
getPivotTableByName(), and replace an inline array_map closure with a typed
helper so PHPStan (level 8) is satisfied.
Shorten @see references to unqualified class names and drop a stray blank
line after the sample header docblock, matching the project's php-cs-fixer
and phpcs rules.
Adds runnable samples under samples/PivotTable demonstrating the pivot table
feature, discovered automatically by the sample browser:
- 01_Create_PivotTable: build a basic pivot (row/column/value fields) from a
source range and save it as Xlsx.
- 02_PivotTable_Page_Filter: add a page (report filter) field.
- 03_PivotTable_Grouping: group a numeric field into ranges and a date field
by quarter.
- 04_Read_PivotTable: write a pivot, load it back, and inspect the pivot table
object model (source, row/column/page fields, value aggregation).
Directory is named PivotTable to match the PascalCase convention used by the
other sample categories (Table, Chart, ...).
Extends PivotTableBuilder with the two remaining common pivot capabilities.
Page fields: addPageField() places a field on the page (report filter) axis,
and the writer emits a <pageFields> section (in its correct schema position,
after colFields and before dataFields) with the field marked axis="axisPage".
Grouping: a new Worksheet\PivotTable\PivotFieldGroup value object captures how a
field is grouped, and the builder gains groupFieldByNumericRange() (fixed-width
buckets between a start and end number) and groupFieldByDate() (group a date
field by a calendar unit such as quarters or months). The grouping is written
into the pivot cache definition as a <fieldGroup> with <rangePr> and computed
<groupItems>; with refresh-on-load the spreadsheet application materialises the
buckets when it opens the file.
The generated cache definition, workbook wiring and content types remain a
structurally consistent OPC package, and pivots with page fields or grouping
read back into the object model. Tests cover page-field placement/output,
numeric range grouping, date quarter grouping, and validation of unknown
fields.
Adds the ability to create a new pivot table from a range of source data,
building on the existing read model and round-trip preservation.
New Worksheet\PivotTable\PivotTableBuilder provides a fluent API: point it at a
source worksheet and range (its first row supplies the field names), place
fields on the row/column axes and add value fields with an aggregation
function, then build() produces a PivotTable and registers it on a target
sheet. Source field names are resolved from the header row (RichText/inline
string headers included), and distinct values are collected for the cache
definition's sharedItems.
The model gains what generation needs: PivotTable::isGenerated(), per-field
data-field captions, subtotal-function constants on PivotField, and sharedItems
on PivotCacheDefinition.
Writer: a new Writer\Xlsx\PivotTable generates the pivotTableDefinition,
pivotCacheDefinition (with refreshOnLoad set) and an empty pivotCacheRecords
part, plus their rels. On save, generated pivot tables are assigned indices
that continue past any preserved pivot parts (so a workbook can mix loaded and
new pivots without collision) and are wired into the worksheet relationships,
workbook <pivotCaches> registry, workbook relationships and content types via
the same unparsed-data path used for preservation. Values are computed by the
spreadsheet application on open via refresh-on-load, so we don't reimplement
the aggregation engine.
Output is a structurally consistent OPC package and reads back into the object
model. Tests cover building, validation errors, the generated parts and
wiring, aggregation attributes, and mixing preserved with generated pivots.
Previously, loading an Xlsx file that contained a pivot table and saving it
again silently dropped every pivot part, corrupting the user's pivot tables.
This makes the writer round-trip them intact.
Reader: alongside parsing pivot tables into the object model, the raw pivot
parts are now captured into the unparsed loaded data - the pivot table part
and its rels (per sheet), the pivot cache definition, its rels and the cache
records - each keyed by its original archive path so the relationships between
them stay valid. The workbook <pivotCaches> registry (cacheId -> cache
definition) and the pivot content-type overrides are captured too.
Writer: the preserved parts are re-added to the archive; worksheet
relationships gain a pivotTable relationship, workbook relationships gain a
pivotCacheDefinition relationship, and workbook.xml re-emits <pivotCaches>
with an r:id that matches the regenerated workbook relationship. Content types
are restored via the existing override_content_types pass-through.
The output is a structurally consistent OPC package (all rels and content
types resolve) and is stable across repeated round-trips. Tests cover part
preservation, workbook/rels/content-type wiring, and that the pivot object
model still reads back after a save.
Introduces a first-class, read-only representation of pivot tables that
already exist in a loaded Xlsx file (issue #4534). Previously pivot tables
were completely inaccessible from the object model.
New model classes under Worksheet\PivotTable:
- PivotTable: name, location, source cache definition and field layout,
with axis helpers (getRowFields/getColumnFields/getPageFields/getDataFields).
- PivotCacheDefinition: cache id, source worksheet/range and cache field names.
- PivotField: field index, name, axis placement, and data-field aggregation.
Reader\Xlsx\PivotTableReader parses a pivotTableDefinition part and its
associated pivotCacheDefinition part into that model. The main Xlsx reader
discovers pivot parts via worksheet relationships and wires them up (skipped
under setReadDataOnly). Worksheet gains getPivotTableCollection(),
getPivotTables(), addPivotTable(), getPivotTableByName(), getPivotTableNames()
and removePivotTableCollection().
This is inspection-only: it does not create, modify, recalculate or render
pivot tables, and does not change what the writer emits. A minimal pivot-table
fixture and functional tests cover reading name, location, cache source,
fields and axis placement, plus the read-data-only path.
GammaBase::calculateInverse fixed its upper bound at alpha*beta*5, so any
GAMMA.INV/GAMMAINV quantile larger than that was clamped to it: e.g.
GAMMAINV(0.9999, 1, 1) returned 5 rather than ~9.2103, breaking the
round-trip GAMMADIST(GAMMAINV(p)) == p.
Grow the upper bound geometrically until it brackets the root. If the CDF
stops increasing first (the series approximation is past its usable range)
keep the original bound instead of expanding into it, which also stops a
probability the series cannot reach from running the bound away.
Per review feedback on PR #4943:
- Remove the $this->ns property from Reader/Xlsx/Sparklines and the
getNamespaces() call that populated it. The reader now navigates the XML
with the fixed namespace URIs (Namespaces::DATA_VALIDATIONS1 for x14 and
Namespaces::DATA_VALIDATIONS2 for xm) via SimpleXMLElement::children(),
so it no longer depends on the prefixes a third-party writer happens to
use.
- Move the SPARKLINE_URI constant to the Namespaces class (alongside
STYLE_CHECKBOX_URI) and reference it from both the reader and writer.
Add tests for previously-uncovered SparklineGroup setters (axis types,
manual limits, additional display options and colours) and a hand-crafted
Xlsx fixture exercising the reader's edge cases (non-sparkline ext, group
without sparklines, sparkline with empty sqref, missing colour elements).
Harden Reader/Xlsx/Sparklines against malformed input: guard the
sparklineGroup and sparkline iterations against a null children set so an
empty <x14:sparklineGroups> or <x14:sparklines> element no longer emits a
PHP warning.
Implements native sparkline support (line, column, and win/loss) in the
Xlsx reader and writer, resolving issue #4941.
- New model classes under Worksheet/Sparkline: Sparkline, SparklineGroup,
and the SparklineType enum.
- Worksheet gains a sparkline group collection with addSparkline(),
addSparklineGroup(), getSparklineGroupCollection(), and
removeSparklineGroupCollection(), plus deep-clone support.
- Reader/Xlsx/Sparklines parses x14:sparklineGroups from the sheet extLst.
- Writer/Xlsx/Worksheet emits x14:sparklineGroups; writeExtLst is
refactored so conditional-formatting data bars and sparklines share a
single extLst.
- Adds unit tests, an Xlsx round-trip test, a runnable sample, and docs.
Julien Chavée
oleibman
dependabot[bot]
Vincent Gao
saifulislamferoz