table.go 15.2 KB
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516
// Copyright 2016 - 2020 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 XLSX / XLSM / XLTM files. Supports reading and writing
// spreadsheet documents generated by Microsoft Exce™ 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 1.10 or later.

package excelize

import (
	"encoding/json"
	"encoding/xml"
	"fmt"
	"regexp"
	"strconv"
	"strings"
)

// parseFormatTableSet provides a function to parse the format settings of the
// table with default value.
func parseFormatTableSet(formatSet string) (*formatTable, error) {
	format := formatTable{
		TableStyle:     "",
		ShowRowStripes: true,
	}
	err := json.Unmarshal(parseFormatSet(formatSet), &format)
	return &format, err
}

// AddTable provides the method to add table in a worksheet by given worksheet
// name, coordinate area and format set. For example, create a table of A1:D5
// on Sheet1:
//
//    err := f.AddTable("Sheet1", "A1", "D5", ``)
//
// Create a table of F2:H6 on Sheet2 with format set:
//
//    err := f.AddTable("Sheet2", "F2", "H6", `{"table_name":"table","table_style":"TableStyleMedium2", "show_first_column":true,"show_last_column":true,"show_row_stripes":false,"show_column_stripes":true}`)
//
// Note that the table must be at least two lines including the header. The
// header cells must contain strings and must be unique, and must set the
// header row data of the table before calling the AddTable function. Multiple
// tables coordinate areas that can't have an intersection.
//
// table_name: The name of the table, in the same worksheet name of the table should be unique
//
// table_style: The built-in table style names
//
//    TableStyleLight1 - TableStyleLight21
//    TableStyleMedium1 - TableStyleMedium28
//    TableStyleDark1 - TableStyleDark11
//
func (f *File) AddTable(sheet, hcell, vcell, format string) error {
	formatSet, err := parseFormatTableSet(format)
	if err != nil {
		return err
	}
	// Coordinate conversion, convert C1:B3 to 2,0,1,2.
	hcol, hrow, err := CellNameToCoordinates(hcell)
	if err != nil {
		return err
	}
	vcol, vrow, err := CellNameToCoordinates(vcell)
	if err != nil {
		return err
	}

	if vcol < hcol {
		vcol, hcol = hcol, vcol
	}

	if vrow < hrow {
		vrow, hrow = hrow, vrow
	}

	tableID := f.countTables() + 1
	sheetRelationshipsTableXML := "../tables/table" + strconv.Itoa(tableID) + ".xml"
	tableXML := strings.Replace(sheetRelationshipsTableXML, "..", "xl", -1)
	// Add first table for given sheet.
	sheetRels := "xl/worksheets/_rels/" + strings.TrimPrefix(f.sheetMap[trimSheetName(sheet)], "xl/worksheets/") + ".rels"
	rID := f.addRels(sheetRels, SourceRelationshipTable, sheetRelationshipsTableXML, "")
	if err = f.addSheetTable(sheet, rID); err != nil {
		return err
	}
	f.addSheetNameSpace(sheet, SourceRelationship)
	if err = f.addTable(sheet, tableXML, hcol, hrow, vcol, vrow, tableID, formatSet); err != nil {
		return err
	}
	f.addContentTypePart(tableID, "table")
	return err
}

// countTables provides a function to get table files count storage in the
// folder xl/tables.
func (f *File) countTables() int {
	count := 0
	for k := range f.XLSX {
		if strings.Contains(k, "xl/tables/table") {
			count++
		}
	}
	return count
}

// addSheetTable provides a function to add tablePart element to
// xl/worksheets/sheet%d.xml by given worksheet name and relationship index.
func (f *File) addSheetTable(sheet string, rID int) error {
	ws, err := f.workSheetReader(sheet)
	if err != nil {
		return err
	}
	table := &xlsxTablePart{
		RID: "rId" + strconv.Itoa(rID),
	}
	if ws.TableParts == nil {
		ws.TableParts = &xlsxTableParts{}
	}
	ws.TableParts.Count++
	ws.TableParts.TableParts = append(ws.TableParts.TableParts, table)
	return err
}

// addTable provides a function to add table by given worksheet name,
// coordinate area and format set.
func (f *File) addTable(sheet, tableXML string, x1, y1, x2, y2, i int, formatSet *formatTable) error {
	// Correct the minimum number of rows, the table at least two lines.
	if y1 == y2 {
		y2++
	}

	// Correct table reference coordinate area, such correct C1:B3 to B1:C3.
	ref, err := f.coordinatesToAreaRef([]int{x1, y1, x2, y2})
	if err != nil {
		return err
	}

	var tableColumn []*xlsxTableColumn

	idx := 0
	for i := x1; i <= x2; i++ {
		idx++
		cell, err := CoordinatesToCellName(i, y1)
		if err != nil {
			return err
		}
		name, _ := f.GetCellValue(sheet, cell)
		if _, err := strconv.Atoi(name); err == nil {
			_ = f.SetCellStr(sheet, cell, name)
		}
		if name == "" {
			name = "Column" + strconv.Itoa(idx)
			f.SetCellStr(sheet, cell, name)
		}
		tableColumn = append(tableColumn, &xlsxTableColumn{
			ID:   idx,
			Name: name,
		})
	}
	name := formatSet.TableName
	if name == "" {
		name = "Table" + strconv.Itoa(i)
	}
	t := xlsxTable{
		XMLNS:       NameSpaceSpreadSheet.Value,
		ID:          i,
		Name:        name,
		DisplayName: name,
		Ref:         ref,
		AutoFilter: &xlsxAutoFilter{
			Ref: ref,
		},
		TableColumns: &xlsxTableColumns{
			Count:       idx,
			TableColumn: tableColumn,
		},
		TableStyleInfo: &xlsxTableStyleInfo{
			Name:              formatSet.TableStyle,
			ShowFirstColumn:   formatSet.ShowFirstColumn,
			ShowLastColumn:    formatSet.ShowLastColumn,
			ShowRowStripes:    formatSet.ShowRowStripes,
			ShowColumnStripes: formatSet.ShowColumnStripes,
		},
	}
	table, _ := xml.Marshal(t)
	f.saveFileList(tableXML, table)
	return nil
}

// parseAutoFilterSet provides a function to parse the settings of the auto
// filter.
func parseAutoFilterSet(formatSet string) (*formatAutoFilter, error) {
	format := formatAutoFilter{}
	err := json.Unmarshal([]byte(formatSet), &format)
	return &format, err
}

// AutoFilter provides the method to add auto filter in a worksheet by given
// worksheet name, coordinate area and settings. An autofilter in Excel is a
// way of filtering a 2D range of data based on some simple criteria. For
// example applying an autofilter to a cell range A1:D4 in the Sheet1:
//
//    err := f.AutoFilter("Sheet1", "A1", "D4", "")
//
// Filter data in an autofilter:
//
//    err := f.AutoFilter("Sheet1", "A1", "D4", `{"column":"B","expression":"x != blanks"}`)
//
// column defines the filter columns in a autofilter range based on simple
// criteria
//
// It isn't sufficient to just specify the filter condition. You must also
// hide any rows that don't match the filter condition. Rows are hidden using
// the SetRowVisible() method. Excelize can't filter rows automatically since
// this isn't part of the file format.
//
// Setting a filter criteria for a column:
//
// expression defines the conditions, the following operators are available
// for setting the filter criteria:
//
//    ==
//    !=
//    >
//    <
//    >=
//    <=
//    and
//    or
//
// An expression can comprise a single statement or two statements separated
// by the 'and' and 'or' operators. For example:
//
//    x <  2000
//    x >  2000
//    x == 2000
//    x >  2000 and x <  5000
//    x == 2000 or  x == 5000
//
// Filtering of blank or non-blank data can be achieved by using a value of
// Blanks or NonBlanks in the expression:
//
//    x == Blanks
//    x == NonBlanks
//
// Excel also allows some simple string matching operations:
//
//    x == b*      // begins with b
//    x != b*      // doesnt begin with b
//    x == *b      // ends with b
//    x != *b      // doesnt end with b
//    x == *b*     // contains b
//    x != *b*     // doesn't contains b
//
// You can also use '*' to match any character or number and '?' to match any
// single character or number. No other regular expression quantifier is
// supported by Excel's filters. Excel's regular expression characters can be
// escaped using '~'.
//
// The placeholder variable x in the above examples can be replaced by any
// simple string. The actual placeholder name is ignored internally so the
// following are all equivalent:
//
//    x     < 2000
//    col   < 2000
//    Price < 2000
//
func (f *File) AutoFilter(sheet, hcell, vcell, format string) error {
	hcol, hrow, err := CellNameToCoordinates(hcell)
	if err != nil {
		return err
	}
	vcol, vrow, err := CellNameToCoordinates(vcell)
	if err != nil {
		return err
	}

	if vcol < hcol {
		vcol, hcol = hcol, vcol
	}

	if vrow < hrow {
		vrow, hrow = hrow, vrow
	}

	formatSet, _ := parseAutoFilterSet(format)
	cellStart, _ := CoordinatesToCellName(hcol, hrow)
	cellEnd, _ := CoordinatesToCellName(vcol, vrow)
	ref, filterDB := cellStart+":"+cellEnd, "_xlnm._FilterDatabase"
	wb := f.workbookReader()
	sheetID := f.GetSheetIndex(sheet)
	filterRange := fmt.Sprintf("%s!%s", sheet, ref)
	d := xlsxDefinedName{
		Name:         filterDB,
		Hidden:       true,
		LocalSheetID: intPtr(sheetID),
		Data:         filterRange,
	}
	if wb.DefinedNames == nil {
		wb.DefinedNames = &xlsxDefinedNames{
			DefinedName: []xlsxDefinedName{d},
		}
	} else {
		var definedNameExists bool
		for idx := range wb.DefinedNames.DefinedName {
			definedName := wb.DefinedNames.DefinedName[idx]
			if definedName.Name == filterDB && *definedName.LocalSheetID == sheetID && definedName.Hidden {
				wb.DefinedNames.DefinedName[idx].Data = filterRange
				definedNameExists = true
			}
		}
		if !definedNameExists {
			wb.DefinedNames.DefinedName = append(wb.DefinedNames.DefinedName, d)
		}
	}
	refRange := vcol - hcol
	return f.autoFilter(sheet, ref, refRange, hcol, formatSet)
}

// autoFilter provides a function to extract the tokens from the filter
// expression. The tokens are mainly non-whitespace groups.
func (f *File) autoFilter(sheet, ref string, refRange, col int, formatSet *formatAutoFilter) error {
	xlsx, err := f.workSheetReader(sheet)
	if err != nil {
		return err
	}
	if xlsx.SheetPr != nil {
		xlsx.SheetPr.FilterMode = true
	}
	xlsx.SheetPr = &xlsxSheetPr{FilterMode: true}
	filter := &xlsxAutoFilter{
		Ref: ref,
	}
	xlsx.AutoFilter = filter
	if formatSet.Column == "" || formatSet.Expression == "" {
		return nil
	}

	fsCol, err := ColumnNameToNumber(formatSet.Column)
	if err != nil {
		return err
	}
	offset := fsCol - col
	if offset < 0 || offset > refRange {
		return fmt.Errorf("incorrect index of column '%s'", formatSet.Column)
	}

	filter.FilterColumn = &xlsxFilterColumn{
		ColID: offset,
	}
	re := regexp.MustCompile(`"(?:[^"]|"")*"|\S+`)
	token := re.FindAllString(formatSet.Expression, -1)
	if len(token) != 3 && len(token) != 7 {
		return fmt.Errorf("incorrect number of tokens in criteria '%s'", formatSet.Expression)
	}
	expressions, tokens, err := f.parseFilterExpression(formatSet.Expression, token)
	if err != nil {
		return err
	}
	f.writeAutoFilter(filter, expressions, tokens)
	xlsx.AutoFilter = filter
	return nil
}

// writeAutoFilter provides a function to check for single or double custom
// filters as default filters and handle them accordingly.
func (f *File) writeAutoFilter(filter *xlsxAutoFilter, exp []int, tokens []string) {
	if len(exp) == 1 && exp[0] == 2 {
		// Single equality.
		var filters []*xlsxFilter
		filters = append(filters, &xlsxFilter{Val: tokens[0]})
		filter.FilterColumn.Filters = &xlsxFilters{Filter: filters}
	} else if len(exp) == 3 && exp[0] == 2 && exp[1] == 1 && exp[2] == 2 {
		// Double equality with "or" operator.
		filters := []*xlsxFilter{}
		for _, v := range tokens {
			filters = append(filters, &xlsxFilter{Val: v})
		}
		filter.FilterColumn.Filters = &xlsxFilters{Filter: filters}
	} else {
		// Non default custom filter.
		expRel := map[int]int{0: 0, 1: 2}
		andRel := map[int]bool{0: true, 1: false}
		for k, v := range tokens {
			f.writeCustomFilter(filter, exp[expRel[k]], v)
			if k == 1 {
				filter.FilterColumn.CustomFilters.And = andRel[exp[k]]
			}
		}
	}
}

// writeCustomFilter provides a function to write the <customFilter> element.
func (f *File) writeCustomFilter(filter *xlsxAutoFilter, operator int, val string) {
	operators := map[int]string{
		1:  "lessThan",
		2:  "equal",
		3:  "lessThanOrEqual",
		4:  "greaterThan",
		5:  "notEqual",
		6:  "greaterThanOrEqual",
		22: "equal",
	}
	customFilter := xlsxCustomFilter{
		Operator: operators[operator],
		Val:      val,
	}
	if filter.FilterColumn.CustomFilters != nil {
		filter.FilterColumn.CustomFilters.CustomFilter = append(filter.FilterColumn.CustomFilters.CustomFilter, &customFilter)
	} else {
		customFilters := []*xlsxCustomFilter{}
		customFilters = append(customFilters, &customFilter)
		filter.FilterColumn.CustomFilters = &xlsxCustomFilters{CustomFilter: customFilters}
	}
}

// parseFilterExpression provides a function to converts the tokens of a
// possibly conditional expression into 1 or 2 sub expressions for further
// parsing.
//
// Examples:
//
//    ('x', '==', 2000) -> exp1
//    ('x', '>',  2000, 'and', 'x', '<', 5000) -> exp1 and exp2
//
func (f *File) parseFilterExpression(expression string, tokens []string) ([]int, []string, error) {
	expressions := []int{}
	t := []string{}
	if len(tokens) == 7 {
		// The number of tokens will be either 3 (for 1 expression) or 7 (for 2
		// expressions).
		conditional := 0
		c := tokens[3]
		re, _ := regexp.Match(`(or|\|\|)`, []byte(c))
		if re {
			conditional = 1
		}
		expression1, token1, err := f.parseFilterTokens(expression, tokens[0:3])
		if err != nil {
			return expressions, t, err
		}
		expression2, token2, err := f.parseFilterTokens(expression, tokens[4:7])
		if err != nil {
			return expressions, t, err
		}
		expressions = []int{expression1[0], conditional, expression2[0]}
		t = []string{token1, token2}
	} else {
		exp, token, err := f.parseFilterTokens(expression, tokens)
		if err != nil {
			return expressions, t, err
		}
		expressions = exp
		t = []string{token}
	}
	return expressions, t, nil
}

// parseFilterTokens provides a function to parse the 3 tokens of a filter
// expression and return the operator and token.
func (f *File) parseFilterTokens(expression string, tokens []string) ([]int, string, error) {
	operators := map[string]int{
		"==": 2,
		"=":  2,
		"=~": 2,
		"eq": 2,
		"!=": 5,
		"!~": 5,
		"ne": 5,
		"<>": 5,
		"<":  1,
		"<=": 3,
		">":  4,
		">=": 6,
	}
	operator, ok := operators[strings.ToLower(tokens[1])]
	if !ok {
		// Convert the operator from a number to a descriptive string.
		return []int{}, "", fmt.Errorf("unknown operator: %s", tokens[1])
	}
	token := tokens[2]
	// Special handling for Blanks/NonBlanks.
	re, _ := regexp.Match("blanks|nonblanks", []byte(strings.ToLower(token)))
	if re {
		// Only allow Equals or NotEqual in this context.
		if operator != 2 && operator != 5 {
			return []int{operator}, token, fmt.Errorf("the operator '%s' in expression '%s' is not valid in relation to Blanks/NonBlanks'", tokens[1], expression)
		}
		token = strings.ToLower(token)
		// The operator should always be 2 (=) to flag a "simple" equality in
		// the binary record. Therefore we convert <> to =.
		if token == "blanks" {
			if operator == 5 {
				token = " "
			}
		} else {
			if operator == 5 {
				operator = 2
				token = "blanks"
			} else {
				operator = 5
				token = " "
			}
		}
	}
	// if the string token contains an Excel match character then change the
	// operator type to indicate a non "simple" equality.
	re, _ = regexp.Match("[*?]", []byte(token))
	if operator == 2 && re {
		operator = 22
	}
	return []int{operator}, token, nil
}