excelize/xmlPivotCache.go

229 lines
11 KiB
Go
Raw Permalink Normal View History

// Copyright 2016 - 2023 The excelize Authors. All rights reserved. Use of
// this source code is governed by a BSD-style license that can be found in
// the LICENSE file.
//
// Package excelize providing a set of functions that allow you to write to and
// read from XLAM / XLSM / XLSX / XLTM / XLTX files. Supports reading and
// writing spreadsheet documents generated by Microsoft Excel™ 2007 and later.
// Supports complex components by high compatibility, and provided streaming
// API for generating or reading data from a worksheet with huge amounts of
// data. This library needs Go version 1.16 or later.
package excelize
import "encoding/xml"
// xlsxPivotCacheDefinition represents the pivotCacheDefinition part. This part
// defines each field in the source data, including the name, the string
// resources of the instance data (for shared items), and information about
// the type of data that appears in the field.
type xlsxPivotCacheDefinition struct {
XMLName xml.Name `xml:"http://schemas.openxmlformats.org/spreadsheetml/2006/main pivotCacheDefinition"`
RID string `xml:"http://schemas.openxmlformats.org/officeDocument/2006/relationships id,attr,omitempty"`
Invalid bool `xml:"invalid,attr,omitempty"`
2019-09-20 00:20:30 +08:00
SaveData bool `xml:"saveData,attr"`
RefreshOnLoad bool `xml:"refreshOnLoad,attr,omitempty"`
OptimizeMemory bool `xml:"optimizeMemory,attr,omitempty"`
EnableRefresh bool `xml:"enableRefresh,attr,omitempty"`
RefreshedBy string `xml:"refreshedBy,attr,omitempty"`
RefreshedDate float64 `xml:"refreshedDate,attr,omitempty"`
RefreshedDateIso float64 `xml:"refreshedDateIso,attr,omitempty"`
BackgroundQuery bool `xml:"backgroundQuery,attr"`
MissingItemsLimit int `xml:"missingItemsLimit,attr,omitempty"`
CreatedVersion int `xml:"createdVersion,attr,omitempty"`
RefreshedVersion int `xml:"refreshedVersion,attr,omitempty"`
MinRefreshableVersion int `xml:"minRefreshableVersion,attr,omitempty"`
RecordCount int `xml:"recordCount,attr,omitempty"`
UpgradeOnRefresh bool `xml:"upgradeOnRefresh,attr,omitempty"`
TupleCacheAttr bool `xml:"tupleCache,attr,omitempty"`
SupportSubquery bool `xml:"supportSubquery,attr,omitempty"`
SupportAdvancedDrill bool `xml:"supportAdvancedDrill,attr,omitempty"`
CacheSource *xlsxCacheSource `xml:"cacheSource"`
CacheFields *xlsxCacheFields `xml:"cacheFields"`
CacheHierarchies *xlsxCacheHierarchies `xml:"cacheHierarchies"`
Kpis *xlsxKpis `xml:"kpis"`
TupleCache *xlsxTupleCache `xml:"tupleCache"`
CalculatedItems *xlsxCalculatedItems `xml:"calculatedItems"`
CalculatedMembers *xlsxCalculatedMembers `xml:"calculatedMembers"`
Dimensions *xlsxDimensions `xml:"dimensions"`
MeasureGroups *xlsxMeasureGroups `xml:"measureGroups"`
Maps *xlsxMaps `xml:"maps"`
ExtLst *xlsxExtLst `xml:"extLst"`
}
// xlsxCacheSource represents the description of data source whose data is
// stored in the pivot cache. The data source refers to the underlying rows or
// database records that provide the data for a PivotTable. You can create a
// PivotTable report from a SpreadsheetML table, an external database
// (including OLAP cubes), multiple SpreadsheetML worksheets, or another
// PivotTable.
type xlsxCacheSource struct {
2019-09-20 00:20:30 +08:00
Type string `xml:"type,attr"`
ConnectionID int `xml:"connectionId,attr,omitempty"`
2019-09-20 00:20:30 +08:00
WorksheetSource *xlsxWorksheetSource `xml:"worksheetSource"`
Consolidation *xlsxConsolidation `xml:"consolidation"`
ExtLst *xlsxExtLst `xml:"extLst"`
}
// xlsxWorksheetSource represents the location of the source of the data that
// is stored in the cache.
type xlsxWorksheetSource struct {
RID string `xml:"http://schemas.openxmlformats.org/officeDocument/2006/relationships id,attr,omitempty"`
Ref string `xml:"ref,attr,omitempty"`
Name string `xml:"name,attr,omitempty"`
Sheet string `xml:"sheet,attr,omitempty"`
}
// xlsxConsolidation represents the description of the PivotCache source using
// multiple consolidation ranges. This element is used when the source of the
// PivotTable is a collection of ranges in the workbook. The ranges are
// specified in the rangeSets collection. The logic for how the application
// consolidates the data in the ranges is application- defined.
type xlsxConsolidation struct{}
// xlsxCacheFields represents the collection of field definitions in the
// source data.
type xlsxCacheFields struct {
Count int `xml:"count,attr"`
CacheField []*xlsxCacheField `xml:"cacheField"`
}
// xlsxCacheField represent a single field in the PivotCache. This definition
// contains information about the field, such as its source, data type, and
// location within a level or hierarchy. The sharedItems element stores
// additional information about the data in this field. If there are no shared
// items, then values are stored directly in the pivotCacheRecords part.
type xlsxCacheField struct {
Name string `xml:"name,attr"`
Caption string `xml:"caption,attr,omitempty"`
PropertyName string `xml:"propertyName,attr,omitempty"`
ServerField bool `xml:"serverField,attr,omitempty"`
UniqueList bool `xml:"uniqueList,attr,omitempty"`
NumFmtID int `xml:"numFmtId,attr"`
Formula string `xml:"formula,attr,omitempty"`
SQLType int `xml:"sqlType,attr,omitempty"`
Hierarchy int `xml:"hierarchy,attr,omitempty"`
Level int `xml:"level,attr,omitempty"`
2019-09-20 00:20:30 +08:00
DatabaseField bool `xml:"databaseField,attr,omitempty"`
MappingCount int `xml:"mappingCount,attr,omitempty"`
MemberPropertyField bool `xml:"memberPropertyField,attr,omitempty"`
SharedItems *xlsxSharedItems `xml:"sharedItems"`
FieldGroup *xlsxFieldGroup `xml:"fieldGroup"`
2019-09-20 00:20:30 +08:00
MpMap *xlsxX `xml:"mpMap"`
ExtLst *xlsxExtLst `xml:"extLst"`
}
// xlsxSharedItems represents the collection of unique items for a field in
// the PivotCacheDefinition. The sharedItems complex type stores data type and
// formatting information about the data in a field. Items in the
// PivotCacheDefinition can be shared in order to reduce the redundancy of
// those values that are referenced in multiple places across all the
// PivotTable parts.
type xlsxSharedItems struct {
ContainsSemiMixedTypes bool `xml:"containsSemiMixedTypes,attr,omitempty"`
ContainsNonDate bool `xml:"containsNonDate,attr,omitempty"`
ContainsDate bool `xml:"containsDate,attr,omitempty"`
ContainsString bool `xml:"containsString,attr,omitempty"`
ContainsBlank bool `xml:"containsBlank,attr,omitempty"`
ContainsMixedTypes bool `xml:"containsMixedTypes,attr,omitempty"`
ContainsNumber bool `xml:"containsNumber,attr,omitempty"`
ContainsInteger bool `xml:"containsInteger,attr,omitempty"`
MinValue float64 `xml:"minValue,attr,omitempty"`
MaxValue float64 `xml:"maxValue,attr,omitempty"`
MinDate string `xml:"minDate,attr,omitempty"`
MaxDate string `xml:"maxDate,attr,omitempty"`
2019-09-20 00:20:30 +08:00
Count int `xml:"count,attr"`
LongText bool `xml:"longText,attr,omitempty"`
M *xlsxMissing `xml:"m"`
N *xlsxNumber `xml:"n"`
B *xlsxBoolean `xml:"b"`
E *xlsxError `xml:"e"`
S *xlsxString `xml:"s"`
D *xlsxDateTime `xml:"d"`
}
// xlsxMissing represents a value that was not specified.
type xlsxMissing struct{}
// xlsxNumber represents a numeric value in the PivotTable.
type xlsxNumber struct {
V float64 `xml:"v,attr"`
U bool `xml:"u,attr,omitempty"`
F bool `xml:"f,attr,omitempty"`
C string `xml:"c,attr,omitempty"`
Cp int `xml:"cp,attr,omitempty"`
In int `xml:"in,attr,omitempty"`
Bc string `xml:"bc,attr,omitempty"`
Fc string `xml:"fc,attr,omitempty"`
I bool `xml:"i,attr,omitempty"`
Un bool `xml:"un,attr,omitempty"`
St bool `xml:"st,attr,omitempty"`
B bool `xml:"b,attr,omitempty"`
Tpls *xlsxTuples `xml:"tpls"`
X *attrValInt `xml:"x"`
}
// xlsxTuples represents members for the OLAP sheet data entry, also known as
// a tuple.
type xlsxTuples struct{}
// xlsxBoolean represents a boolean value for an item in the PivotTable.
type xlsxBoolean struct{}
// xlsxError represents an error value. The use of this item indicates that an
// error value is present in the PivotTable source. The error is recorded in
// the value attribute.
type xlsxError struct{}
// xlsxString represents a character value in a PivotTable.
type xlsxString struct {
V string `xml:"v,attr"`
U bool `xml:"u,attr,omitempty"`
F bool `xml:"f,attr,omitempty"`
C string `xml:"c,attr,omitempty"`
Cp int `xml:"cp,attr,omitempty"`
In int `xml:"in,attr,omitempty"`
Bc string `xml:"bc,attr,omitempty"`
Fc string `xml:"fc,attr,omitempty"`
I bool `xml:"i,attr,omitempty"`
Un bool `xml:"un,attr,omitempty"`
St bool `xml:"st,attr,omitempty"`
B bool `xml:"b,attr,omitempty"`
Tpls *xlsxTuples `xml:"tpls"`
X *attrValInt `xml:"x"`
}
// xlsxDateTime represents a date-time value in the PivotTable.
type xlsxDateTime struct{}
// xlsxFieldGroup represents the collection of properties for a field group.
type xlsxFieldGroup struct{}
// xlsxCacheHierarchies represents the collection of OLAP hierarchies in the
// PivotCache.
type xlsxCacheHierarchies struct{}
// xlsxKpis represents the collection of Key Performance Indicators (KPIs)
// defined on the OLAP server and stored in the PivotCache.
type xlsxKpis struct{}
// xlsxTupleCache represents the cache of OLAP sheet data members, or tuples.
type xlsxTupleCache struct{}
// xlsxCalculatedItems represents the collection of calculated items.
type xlsxCalculatedItems struct{}
// xlsxCalculatedMembers represents the collection of calculated members in an
// OLAP PivotTable.
type xlsxCalculatedMembers struct{}
// xlsxDimensions represents the collection of PivotTable OLAP dimensions.
type xlsxDimensions struct{}
// xlsxMeasureGroups represents the collection of PivotTable OLAP measure
// groups.
type xlsxMeasureGroups struct{}
// xlsxMaps represents the PivotTable OLAP measure group - Dimension maps.
type xlsxMaps struct{}