// Copyright 2016 - 2022 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 Excelâ„¢ 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 1.15 or later. package excelize import ( "bytes" "encoding/xml" "fmt" "io" "log" "math" "os" "strconv" "github.com/mohae/deepcopy" ) // GetRows return all the rows in a sheet by given worksheet name, returned as // a two-dimensional array, where the value of the cell is converted to the // string type. If the cell format can be applied to the value of the cell, // the applied value will be used, otherwise the original value will be used. // GetRows fetched the rows with value or formula cells, the continually blank // cells in the tail of each row will be skipped, so the length of each row // may be inconsistent. // // For example, get and traverse the value of all cells by rows on a worksheet // named 'Sheet1': // // rows, err := f.GetRows("Sheet1") // if err != nil { // fmt.Println(err) // return // } // for _, row := range rows { // for _, colCell := range row { // fmt.Print(colCell, "\t") // } // fmt.Println() // } func (f *File) GetRows(sheet string, opts ...Options) ([][]string, error) { rows, err := f.Rows(sheet) if err != nil { return nil, err } results, cur, max := make([][]string, 0, 64), 0, 0 for rows.Next() { cur++ row, err := rows.Columns(opts...) if err != nil { break } results = append(results, row) if len(row) > 0 { max = cur } } return results[:max], rows.Close() } // Rows defines an iterator to a sheet. type Rows struct { err error curRow, seekRow int needClose, rawCellValue bool sheet string f *File tempFile *os.File sst *xlsxSST decoder *xml.Decoder token xml.Token curRowOpts, seekRowOpts RowOpts } // Next will return true if find the next row element. func (rows *Rows) Next() bool { rows.seekRow++ if rows.curRow >= rows.seekRow { rows.curRowOpts = rows.seekRowOpts return true } for { token, _ := rows.decoder.Token() if token == nil { return false } switch xmlElement := token.(type) { case xml.StartElement: if xmlElement.Name.Local == "row" { rows.curRow++ if rowNum, _ := attrValToInt("r", xmlElement.Attr); rowNum != 0 { rows.curRow = rowNum } rows.token = token rows.curRowOpts = extractRowOpts(xmlElement.Attr) return true } case xml.EndElement: if xmlElement.Name.Local == "sheetData" { return false } } } } // GetRowOpts will return the RowOpts of the current row. func (rows *Rows) GetRowOpts() RowOpts { return rows.curRowOpts } // Error will return the error when the error occurs. func (rows *Rows) Error() error { return rows.err } // Close closes the open worksheet XML file in the system temporary // directory. func (rows *Rows) Close() error { if rows.tempFile != nil { return rows.tempFile.Close() } return nil } // Columns return the current row's column values. This fetches the worksheet // data as a stream, returns each cell in a row as is, and will not skip empty // rows in the tail of the worksheet. func (rows *Rows) Columns(opts ...Options) ([]string, error) { if rows.curRow > rows.seekRow { return nil, nil } var rowIterator rowXMLIterator var token xml.Token rows.rawCellValue, rows.sst = parseOptions(opts...).RawCellValue, rows.f.sharedStringsReader() for { if rows.token != nil { token = rows.token } else if token, _ = rows.decoder.Token(); token == nil { break } switch xmlElement := token.(type) { case xml.StartElement: rowIterator.inElement = xmlElement.Name.Local if rowIterator.inElement == "row" { rowNum := 0 if rowNum, rowIterator.err = attrValToInt("r", xmlElement.Attr); rowNum != 0 { rows.curRow = rowNum } else if rows.token == nil { rows.curRow++ } rows.token = token rows.seekRowOpts = extractRowOpts(xmlElement.Attr) if rows.curRow > rows.seekRow { rows.token = nil return rowIterator.columns, rowIterator.err } } if rows.rowXMLHandler(&rowIterator, &xmlElement, rows.rawCellValue); rowIterator.err != nil { rows.token = nil return rowIterator.columns, rowIterator.err } rows.token = nil case xml.EndElement: if xmlElement.Name.Local == "sheetData" { return rowIterator.columns, rowIterator.err } } } return rowIterator.columns, rowIterator.err } // extractRowOpts extract row element attributes. func extractRowOpts(attrs []xml.Attr) RowOpts { rowOpts := RowOpts{Height: defaultRowHeight} if styleID, err := attrValToInt("s", attrs); err == nil && styleID > 0 && styleID < MaxCellStyles { rowOpts.StyleID = styleID } if hidden, err := attrValToBool("hidden", attrs); err == nil { rowOpts.Hidden = hidden } if height, err := attrValToFloat("ht", attrs); err == nil { rowOpts.Height = height } return rowOpts } // appendSpace append blank characters to slice by given length and source slice. func appendSpace(l int, s []string) []string { for i := 1; i < l; i++ { s = append(s, "") } return s } // ErrSheetNotExist defines an error of sheet that does not exist type ErrSheetNotExist struct { SheetName string } func (err ErrSheetNotExist) Error() string { return fmt.Sprintf("sheet %s does not exist", err.SheetName) } // rowXMLIterator defined runtime use field for the worksheet row SAX parser. type rowXMLIterator struct { err error inElement string cellCol int columns []string } // rowXMLHandler parse the row XML element of the worksheet. func (rows *Rows) rowXMLHandler(rowIterator *rowXMLIterator, xmlElement *xml.StartElement, raw bool) { if rowIterator.inElement == "c" { rowIterator.cellCol++ colCell := xlsxC{} _ = rows.decoder.DecodeElement(&colCell, xmlElement) if colCell.R != "" { if rowIterator.cellCol, _, rowIterator.err = CellNameToCoordinates(colCell.R); rowIterator.err != nil { return } } blank := rowIterator.cellCol - len(rowIterator.columns) if val, _ := colCell.getValueFrom(rows.f, rows.sst, raw); val != "" || colCell.F != nil { rowIterator.columns = append(appendSpace(blank, rowIterator.columns), val) } } } // Rows returns a rows iterator, used for streaming reading data for a // worksheet with a large data. This function is concurrency safe. For // example: // // rows, err := f.Rows("Sheet1") // if err != nil { // fmt.Println(err) // return // } // for rows.Next() { // row, err := rows.Columns() // if err != nil { // fmt.Println(err) // } // for _, colCell := range row { // fmt.Print(colCell, "\t") // } // fmt.Println() // } // if err = rows.Close(); err != nil { // fmt.Println(err) // } func (f *File) Rows(sheet string) (*Rows, error) { name, ok := f.getSheetXMLPath(sheet) if !ok { return nil, ErrSheetNotExist{sheet} } if ws, ok := f.Sheet.Load(name); ok && ws != nil { worksheet := ws.(*xlsxWorksheet) worksheet.Lock() defer worksheet.Unlock() // flush data output, _ := xml.Marshal(worksheet) f.saveFileList(name, f.replaceNameSpaceBytes(name, output)) } var err error rows := Rows{f: f, sheet: name} rows.needClose, rows.decoder, rows.tempFile, err = f.xmlDecoder(name) return &rows, err } // getFromStringItem build shared string item offset list from system temporary // file at one time, and return value by given to string index. func (f *File) getFromStringItem(index int) string { if f.sharedStringTemp != nil { if len(f.sharedStringItem) <= index { return strconv.Itoa(index) } offsetRange := f.sharedStringItem[index] buf := make([]byte, offsetRange[1]-offsetRange[0]) if _, err := f.sharedStringTemp.ReadAt(buf, int64(offsetRange[0])); err != nil { return strconv.Itoa(index) } return string(buf) } needClose, decoder, tempFile, err := f.xmlDecoder(defaultXMLPathSharedStrings) if needClose && err == nil { defer tempFile.Close() } f.sharedStringItem = [][]uint{} f.sharedStringTemp, _ = os.CreateTemp(os.TempDir(), "excelize-") f.tempFiles.Store(defaultTempFileSST, f.sharedStringTemp.Name()) var ( inElement string i, offset uint ) for { token, _ := decoder.Token() if token == nil { break } switch xmlElement := token.(type) { case xml.StartElement: inElement = xmlElement.Name.Local if inElement == "si" { si := xlsxSI{} _ = decoder.DecodeElement(&si, &xmlElement) startIdx := offset n, _ := f.sharedStringTemp.WriteString(si.String()) offset += uint(n) f.sharedStringItem = append(f.sharedStringItem, []uint{startIdx, offset}) i++ } } } return f.getFromStringItem(index) } // xmlDecoder creates XML decoder by given path in the zip from memory data // or system temporary file. func (f *File) xmlDecoder(name string) (bool, *xml.Decoder, *os.File, error) { var ( content []byte err error tempFile *os.File ) if content = f.readXML(name); len(content) > 0 { return false, f.xmlNewDecoder(bytes.NewReader(content)), tempFile, err } tempFile, err = f.readTemp(name) return true, f.xmlNewDecoder(tempFile), tempFile, err } // SetRowHeight provides a function to set the height of a single row. For // example, set the height of the first row in Sheet1: // // err := f.SetRowHeight("Sheet1", 1, 50) func (f *File) SetRowHeight(sheet string, row int, height float64) error { if row < 1 { return newInvalidRowNumberError(row) } if height > MaxRowHeight { return ErrMaxRowHeight } ws, err := f.workSheetReader(sheet) if err != nil { return err } prepareSheetXML(ws, 0, row) rowIdx := row - 1 ws.SheetData.Row[rowIdx].Ht = height ws.SheetData.Row[rowIdx].CustomHeight = true return nil } // getRowHeight provides a function to get row height in pixels by given sheet // name and row number. func (f *File) getRowHeight(sheet string, row int) int { ws, _ := f.workSheetReader(sheet) ws.Lock() defer ws.Unlock() for i := range ws.SheetData.Row { v := &ws.SheetData.Row[i] if v.R == row && v.Ht != 0 { return int(convertRowHeightToPixels(v.Ht)) } } // Optimization for when the row heights haven't changed. return int(defaultRowHeightPixels) } // GetRowHeight provides a function to get row height by given worksheet name // and row number. For example, get the height of the first row in Sheet1: // // height, err := f.GetRowHeight("Sheet1", 1) func (f *File) GetRowHeight(sheet string, row int) (float64, error) { if row < 1 { return defaultRowHeightPixels, newInvalidRowNumberError(row) } ht := defaultRowHeight ws, err := f.workSheetReader(sheet) if err != nil { return ht, err } if ws.SheetFormatPr != nil && ws.SheetFormatPr.CustomHeight { ht = ws.SheetFormatPr.DefaultRowHeight } if row > len(ws.SheetData.Row) { return ht, nil // it will be better to use 0, but we take care with BC } for _, v := range ws.SheetData.Row { if v.R == row && v.Ht != 0 { return v.Ht, nil } } // Optimization for when the row heights haven't changed. return ht, nil } // sharedStringsReader provides a function to get the pointer to the structure // after deserialization of xl/sharedStrings.xml. func (f *File) sharedStringsReader() *xlsxSST { var err error f.Lock() defer f.Unlock() relPath := f.getWorkbookRelsPath() if f.SharedStrings == nil { var sharedStrings xlsxSST ss := f.readXML(defaultXMLPathSharedStrings) if err = f.xmlNewDecoder(bytes.NewReader(namespaceStrictToTransitional(ss))). Decode(&sharedStrings); err != nil && err != io.EOF { log.Printf("xml decode error: %s", err) } if sharedStrings.Count == 0 { sharedStrings.Count = len(sharedStrings.SI) } if sharedStrings.UniqueCount == 0 { sharedStrings.UniqueCount = sharedStrings.Count } f.SharedStrings = &sharedStrings for i := range sharedStrings.SI { if sharedStrings.SI[i].T != nil { f.sharedStringsMap[sharedStrings.SI[i].T.Val] = i } } f.addContentTypePart(0, "sharedStrings") rels := f.relsReader(relPath) for _, rel := range rels.Relationships { if rel.Target == "/xl/sharedStrings.xml" { return f.SharedStrings } } // Update workbook.xml.rels f.addRels(relPath, SourceRelationshipSharedStrings, "/xl/sharedStrings.xml", "") } return f.SharedStrings } // getValueFrom return a value from a column/row cell, this function is // intended to be used with for range on rows an argument with the spreadsheet // opened file. func (c *xlsxC) getValueFrom(f *File, d *xlsxSST, raw bool) (string, error) { f.Lock() defer f.Unlock() switch c.T { case "b": if !raw { if c.V == "1" { return "TRUE", nil } if c.V == "0" { return "FALSE", nil } } return f.formattedValue(c.S, c.V, raw), nil case "s": if c.V != "" { xlsxSI := 0 xlsxSI, _ = strconv.Atoi(c.V) if _, ok := f.tempFiles.Load(defaultXMLPathSharedStrings); ok { return f.formattedValue(c.S, f.getFromStringItem(xlsxSI), raw), nil } if len(d.SI) > xlsxSI { return f.formattedValue(c.S, d.SI[xlsxSI].String(), raw), nil } } return f.formattedValue(c.S, c.V, raw), nil case "str": return f.formattedValue(c.S, c.V, raw), nil case "inlineStr": if c.IS != nil { return f.formattedValue(c.S, c.IS.String(), raw), nil } return f.formattedValue(c.S, c.V, raw), nil default: if isNum, precision, decimal := isNumeric(c.V); isNum && !raw { if precision > 15 { c.V = strconv.FormatFloat(decimal, 'G', 15, 64) } else { c.V = strconv.FormatFloat(decimal, 'f', -1, 64) } } return f.formattedValue(c.S, c.V, raw), nil } } // SetRowVisible provides a function to set visible of a single row by given // worksheet name and Excel row number. For example, hide row 2 in Sheet1: // // err := f.SetRowVisible("Sheet1", 2, false) func (f *File) SetRowVisible(sheet string, row int, visible bool) error { if row < 1 { return newInvalidRowNumberError(row) } ws, err := f.workSheetReader(sheet) if err != nil { return err } prepareSheetXML(ws, 0, row) ws.SheetData.Row[row-1].Hidden = !visible return nil } // GetRowVisible provides a function to get visible of a single row by given // worksheet name and Excel row number. For example, get visible state of row // 2 in Sheet1: // // visible, err := f.GetRowVisible("Sheet1", 2) func (f *File) GetRowVisible(sheet string, row int) (bool, error) { if row < 1 { return false, newInvalidRowNumberError(row) } ws, err := f.workSheetReader(sheet) if err != nil { return false, err } if row > len(ws.SheetData.Row) { return false, nil } return !ws.SheetData.Row[row-1].Hidden, nil } // SetRowOutlineLevel provides a function to set outline level number of a // single row by given worksheet name and Excel row number. The value of // parameter 'level' is 1-7. For example, outline row 2 in Sheet1 to level 1: // // err := f.SetRowOutlineLevel("Sheet1", 2, 1) func (f *File) SetRowOutlineLevel(sheet string, row int, level uint8) error { if row < 1 { return newInvalidRowNumberError(row) } if level > 7 || level < 1 { return ErrOutlineLevel } ws, err := f.workSheetReader(sheet) if err != nil { return err } prepareSheetXML(ws, 0, row) ws.SheetData.Row[row-1].OutlineLevel = level return nil } // GetRowOutlineLevel provides a function to get outline level number of a // single row by given worksheet name and Excel row number. For example, get // outline number of row 2 in Sheet1: // // level, err := f.GetRowOutlineLevel("Sheet1", 2) func (f *File) GetRowOutlineLevel(sheet string, row int) (uint8, error) { if row < 1 { return 0, newInvalidRowNumberError(row) } ws, err := f.workSheetReader(sheet) if err != nil { return 0, err } if row > len(ws.SheetData.Row) { return 0, nil } return ws.SheetData.Row[row-1].OutlineLevel, nil } // RemoveRow provides a function to remove single row by given worksheet name // and Excel row number. For example, remove row 3 in Sheet1: // // err := f.RemoveRow("Sheet1", 3) // // Use this method with caution, which will affect changes in references such // as formulas, charts, and so on. If there is any referenced value of the // worksheet, it will cause a file error when you open it. The excelize only // partially updates these references currently. func (f *File) RemoveRow(sheet string, row int) error { if row < 1 { return newInvalidRowNumberError(row) } ws, err := f.workSheetReader(sheet) if err != nil { return err } if row > len(ws.SheetData.Row) { return f.adjustHelper(sheet, rows, row, -1) } keep := 0 for rowIdx := 0; rowIdx < len(ws.SheetData.Row); rowIdx++ { v := &ws.SheetData.Row[rowIdx] if v.R != row { ws.SheetData.Row[keep] = *v keep++ } } ws.SheetData.Row = ws.SheetData.Row[:keep] return f.adjustHelper(sheet, rows, row, -1) } // InsertRows provides a function to insert new rows after the given Excel row // number starting from 1 and number of rows. For example, create two rows // before row 3 in Sheet1: // // err := f.InsertRows("Sheet1", 3, 2) // // Use this method with caution, which will affect changes in references such // as formulas, charts, and so on. If there is any referenced value of the // worksheet, it will cause a file error when you open it. The excelize only // partially updates these references currently. func (f *File) InsertRows(sheet string, row, n int) error { if row < 1 { return newInvalidRowNumberError(row) } if row >= TotalRows || n >= TotalRows { return ErrMaxRows } if n < 1 { return ErrParameterInvalid } return f.adjustHelper(sheet, rows, row, n) } // DuplicateRow inserts a copy of specified row (by its Excel row number) below // // err := f.DuplicateRow("Sheet1", 2) // // Use this method with caution, which will affect changes in references such // as formulas, charts, and so on. If there is any referenced value of the // worksheet, it will cause a file error when you open it. The excelize only // partially updates these references currently. func (f *File) DuplicateRow(sheet string, row int) error { return f.DuplicateRowTo(sheet, row, row+1) } // DuplicateRowTo inserts a copy of specified row by it Excel number // to specified row position moving down exists rows after target position // // err := f.DuplicateRowTo("Sheet1", 2, 7) // // Use this method with caution, which will affect changes in references such // as formulas, charts, and so on. If there is any referenced value of the // worksheet, it will cause a file error when you open it. The excelize only // partially updates these references currently. func (f *File) DuplicateRowTo(sheet string, row, row2 int) error { if row < 1 { return newInvalidRowNumberError(row) } ws, err := f.workSheetReader(sheet) if err != nil { return err } if row2 < 1 || row == row2 { return nil } var ok bool var rowCopy xlsxRow for i, r := range ws.SheetData.Row { if r.R == row { rowCopy = deepcopy.Copy(ws.SheetData.Row[i]).(xlsxRow) ok = true break } } if err := f.adjustHelper(sheet, rows, row2, 1); err != nil { return err } if !ok { return nil } idx2 := -1 for i, r := range ws.SheetData.Row { if r.R == row2 { idx2 = i break } } if idx2 == -1 && len(ws.SheetData.Row) >= row2 { return nil } rowCopy.C = append(make([]xlsxC, 0, len(rowCopy.C)), rowCopy.C...) f.adjustSingleRowDimensions(&rowCopy, row2) if idx2 != -1 { ws.SheetData.Row[idx2] = rowCopy } else { ws.SheetData.Row = append(ws.SheetData.Row, rowCopy) } return f.duplicateMergeCells(sheet, ws, row, row2) } // duplicateMergeCells merge cells in the destination row if there are single // row merged cells in the copied row. func (f *File) duplicateMergeCells(sheet string, ws *xlsxWorksheet, row, row2 int) error { if ws.MergeCells == nil { return nil } if row > row2 { row++ } for _, rng := range ws.MergeCells.Cells { coordinates, err := rangeRefToCoordinates(rng.Ref) if err != nil { return err } if coordinates[1] < row2 && row2 < coordinates[3] { return nil } } for i := 0; i < len(ws.MergeCells.Cells); i++ { mergedCells := ws.MergeCells.Cells[i] coordinates, _ := rangeRefToCoordinates(mergedCells.Ref) x1, y1, x2, y2 := coordinates[0], coordinates[1], coordinates[2], coordinates[3] if y1 == y2 && y1 == row { from, _ := CoordinatesToCellName(x1, row2) to, _ := CoordinatesToCellName(x2, row2) if err := f.MergeCell(sheet, from, to); err != nil { return err } } } return nil } // checkRow provides a function to check and fill each column element for all // rows and make that is continuous in a worksheet of XML. For example: // // // // // // // // // in this case, we should to change it to // // // // // // // // // // // // Noteice: this method could be very slow for large spreadsheets (more than // 3000 rows one sheet). func checkRow(ws *xlsxWorksheet) error { for rowIdx := range ws.SheetData.Row { rowData := &ws.SheetData.Row[rowIdx] colCount := len(rowData.C) if colCount == 0 { continue } // check and fill the cell without r attribute in a row element rCount := 0 for idx, cell := range rowData.C { rCount++ if cell.R != "" { lastR, _, err := CellNameToCoordinates(cell.R) if err != nil { return err } if lastR > rCount { rCount = lastR } continue } rowData.C[idx].R, _ = CoordinatesToCellName(rCount, rowIdx+1) } lastCol, _, err := CellNameToCoordinates(rowData.C[colCount-1].R) if err != nil { return err } if colCount < lastCol { oldList := rowData.C newlist := make([]xlsxC, 0, lastCol) rowData.C = ws.SheetData.Row[rowIdx].C[:0] for colIdx := 0; colIdx < lastCol; colIdx++ { cellName, err := CoordinatesToCellName(colIdx+1, rowIdx+1) if err != nil { return err } newlist = append(newlist, xlsxC{R: cellName}) } rowData.C = newlist for colIdx := range oldList { colData := &oldList[colIdx] colNum, _, err := CellNameToCoordinates(colData.R) if err != nil { return err } ws.SheetData.Row[rowIdx].C[colNum-1] = *colData } } } return nil } // SetRowStyle provides a function to set the style of rows by given worksheet // name, row range, and style ID. Note that this will overwrite the existing // styles for the rows, it won't append or merge style with existing styles. // // For example set style of row 1 on Sheet1: // // err = f.SetRowStyle("Sheet1", 1, 1, styleID) // // Set style of rows 1 to 10 on Sheet1: // // err = f.SetRowStyle("Sheet1", 1, 10, styleID) func (f *File) SetRowStyle(sheet string, start, end, styleID int) error { if end < start { start, end = end, start } if start < 1 { return newInvalidRowNumberError(start) } if end > TotalRows { return ErrMaxRows } s := f.stylesReader() s.Lock() defer s.Unlock() if styleID < 0 || s.CellXfs == nil || len(s.CellXfs.Xf) <= styleID { return newInvalidStyleID(styleID) } ws, err := f.workSheetReader(sheet) if err != nil { return err } prepareSheetXML(ws, 0, end) for row := start - 1; row < end; row++ { ws.SheetData.Row[row].S = styleID ws.SheetData.Row[row].CustomFormat = true for i := range ws.SheetData.Row[row].C { if _, rowNum, err := CellNameToCoordinates(ws.SheetData.Row[row].C[i].R); err == nil && rowNum-1 == row { ws.SheetData.Row[row].C[i].S = styleID } } } return nil } // convertRowHeightToPixels provides a function to convert the height of a // cell from user's units to pixels. If the height hasn't been set by the user // we use the default value. If the row is hidden it has a value of zero. func convertRowHeightToPixels(height float64) float64 { var pixels float64 if height == 0 { return pixels } pixels = math.Ceil(4.0 / 3.0 * height) return pixels }