// 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 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 1.16 or later. package excelize import ( "bytes" "encoding/xml" "fmt" "io" "os" "reflect" "strconv" "strings" "time" ) // StreamWriter defined the type of stream writer. type StreamWriter struct { file *File Sheet string SheetID int sheetWritten bool cols strings.Builder worksheet *xlsxWorksheet rawData bufferedWriter rows int mergeCellsCount int mergeCells strings.Builder tableParts string } // NewStreamWriter return stream writer struct by given worksheet name for // generate new worksheet with large amounts of data. Note that after set // rows, you must call the 'Flush' method to end the streaming writing process // and ensure that the order of row numbers is ascending, the normal mode // functions and stream mode functions can't be work mixed to writing data on // the worksheets, you can't get cell value when in-memory chunks data over // 16MB. For example, set data for worksheet of size 102400 rows x 50 columns // with numbers and style: // // file := excelize.NewFile() // defer func() { // if err := file.Close(); err != nil { // fmt.Println(err) // } // }() // streamWriter, err := file.NewStreamWriter("Sheet1") // if err != nil { // fmt.Println(err) // return // } // styleID, err := file.NewStyle(&excelize.Style{Font: &excelize.Font{Color: "#777777"}}) // if err != nil { // fmt.Println(err) // return // } // if err := streamWriter.SetRow("A1", // []interface{}{ // excelize.Cell{StyleID: styleID, Value: "Data"}, // []excelize.RichTextRun{ // {Text: "Rich ", Font: &excelize.Font{Color: "2354e8"}}, // {Text: "Text", Font: &excelize.Font{Color: "e83723"}}, // }, // }, // excelize.RowOpts{Height: 45, Hidden: false}); err != nil { // fmt.Println(err) // return // } // for rowID := 2; rowID <= 102400; rowID++ { // row := make([]interface{}, 50) // for colID := 0; colID < 50; colID++ { // row[colID] = rand.Intn(640000) // } // cell, _ := excelize.CoordinatesToCellName(1, rowID) // if err := streamWriter.SetRow(cell, row); err != nil { // fmt.Println(err) // return // } // } // if err := streamWriter.Flush(); err != nil { // fmt.Println(err) // return // } // if err := file.SaveAs("Book1.xlsx"); err != nil { // fmt.Println(err) // } // // Set cell value and cell formula for a worksheet with stream writer: // // err := streamWriter.SetRow("A1", []interface{}{ // excelize.Cell{Value: 1}, // excelize.Cell{Value: 2}, // excelize.Cell{Formula: "SUM(A1,B1)"}}); // // Set cell value and rows style for a worksheet with stream writer: // // err := streamWriter.SetRow("A1", []interface{}{ // excelize.Cell{Value: 1}}, // excelize.RowOpts{StyleID: styleID, Height: 20, Hidden: false}); func (f *File) NewStreamWriter(sheet string) (*StreamWriter, error) { if err := checkSheetName(sheet); err != nil { return nil, err } sheetID := f.getSheetID(sheet) if sheetID == -1 { return nil, newNoExistSheetError(sheet) } sw := &StreamWriter{ file: f, Sheet: sheet, SheetID: sheetID, } var err error sw.worksheet, err = f.workSheetReader(sheet) if err != nil { return nil, err } sheetXMLPath, _ := f.getSheetXMLPath(sheet) if f.streams == nil { f.streams = make(map[string]*StreamWriter) } f.streams[sheetXMLPath] = sw _, _ = sw.rawData.WriteString(xml.Header + ``, rID) if err = sw.file.addContentTypePart(tableID, "table"); err != nil { return err } b, _ := xml.Marshal(table) sw.file.saveFileList(tableXML, b) return err } // Extract values from a row in the StreamWriter. func (sw *StreamWriter) getRowValues(hRow, hCol, vCol int) (res []string, err error) { res = make([]string, vCol-hCol+1) r, err := sw.rawData.Reader() if err != nil { return nil, err } dec := sw.file.xmlNewDecoder(r) for { token, err := dec.Token() if err == io.EOF { return res, nil } if err != nil { return nil, err } startElement, ok := getRowElement(token, hRow) if !ok { continue } // decode cells var row xlsxRow if err := dec.DecodeElement(&row, &startElement); err != nil { return nil, err } for _, c := range row.C { col, _, err := CellNameToCoordinates(c.R) if err != nil { return nil, err } if col < hCol || col > vCol { continue } res[col-hCol], _ = c.getValueFrom(sw.file, nil, false) } return res, nil } } // Check if the token is an XLSX row with the matching row number. func getRowElement(token xml.Token, hRow int) (startElement xml.StartElement, ok bool) { startElement, ok = token.(xml.StartElement) if !ok { return } ok = startElement.Name.Local == "row" if !ok { return } ok = false for _, attr := range startElement.Attr { if attr.Name.Local != "r" { continue } row, _ := strconv.Atoi(attr.Value) if row == hRow { ok = true return } } return } // Cell can be used directly in StreamWriter.SetRow to specify a style and // a value. type Cell struct { StyleID int Formula string Value interface{} } // RowOpts define the options for the set row, it can be used directly in // StreamWriter.SetRow to specify the style and properties of the row. type RowOpts struct { Height float64 Hidden bool StyleID int OutlineLevel int } // marshalAttrs prepare attributes of the row. func (r *RowOpts) marshalAttrs() (strings.Builder, error) { var ( err error attrs strings.Builder ) if r == nil { return attrs, err } if r.Height > MaxRowHeight { err = ErrMaxRowHeight return attrs, err } if r.OutlineLevel > 7 { err = ErrOutlineLevel return attrs, err } if r.StyleID > 0 { attrs.WriteString(` s="`) attrs.WriteString(strconv.Itoa(r.StyleID)) attrs.WriteString(`" customFormat="1"`) } if r.Height > 0 { attrs.WriteString(` ht="`) attrs.WriteString(strconv.FormatFloat(r.Height, 'f', -1, 64)) attrs.WriteString(`" customHeight="1"`) } if r.OutlineLevel > 0 { attrs.WriteString(` outlineLevel="`) attrs.WriteString(strconv.Itoa(r.OutlineLevel)) attrs.WriteString(`"`) } if r.Hidden { attrs.WriteString(` hidden="1"`) } return attrs, err } // parseRowOpts provides a function to parse the optional settings for // *StreamWriter.SetRow. func parseRowOpts(opts ...RowOpts) *RowOpts { options := &RowOpts{} for _, opt := range opts { options = &opt } return options } // SetRow writes an array to stream rows by giving starting cell reference and a // pointer to an array of values. Note that you must call the 'Flush' function // to end the streaming writing process. // // As a special case, if Cell is used as a value, then the Cell.StyleID will be // applied to that cell. func (sw *StreamWriter) SetRow(cell string, values []interface{}, opts ...RowOpts) error { col, row, err := CellNameToCoordinates(cell) if err != nil { return err } if row <= sw.rows { return newStreamSetRowError(row) } sw.rows = row sw.writeSheetData() options := parseRowOpts(opts...) attrs, err := options.marshalAttrs() if err != nil { return err } _, _ = sw.rawData.WriteString(``) for i, val := range values { if val == nil { continue } ref, err := CoordinatesToCellName(col+i, row) if err != nil { return err } c := xlsxC{R: ref, S: options.StyleID} if v, ok := val.(Cell); ok { c.S = v.StyleID val = v.Value setCellFormula(&c, v.Formula) } else if v, ok := val.(*Cell); ok && v != nil { c.S = v.StyleID val = v.Value setCellFormula(&c, v.Formula) } if err = sw.setCellValFunc(&c, val); err != nil { _, _ = sw.rawData.WriteString(``) return err } writeCell(&sw.rawData, c) } _, _ = sw.rawData.WriteString(``) return sw.rawData.Sync() } // SetColWidth provides a function to set the width of a single column or // multiple columns for the StreamWriter. Note that you must call // the 'SetColWidth' function before the 'SetRow' function. For example set // the width column B:C as 20: // // err := streamWriter.SetColWidth(2, 3, 20) func (sw *StreamWriter) SetColWidth(min, max int, width float64) error { if sw.sheetWritten { return ErrStreamSetColWidth } if min < MinColumns || min > MaxColumns || max < MinColumns || max > MaxColumns { return ErrColumnNumber } if width > MaxColumnWidth { return ErrColumnWidth } if min > max { min, max = max, min } sw.cols.WriteString(``) return nil } // InsertPageBreak creates a page break to determine where the printed page ends // and where begins the next one by a given cell reference, the content before // the page break will be printed on one page and after the page break on // another. func (sw *StreamWriter) InsertPageBreak(cell string) error { return sw.worksheet.insertPageBreak(cell) } // SetPanes provides a function to create and remove freeze panes and split // panes by giving panes options for the StreamWriter. Note that you must call // the 'SetPanes' function before the 'SetRow' function. func (sw *StreamWriter) SetPanes(panes *Panes) error { if sw.sheetWritten { return ErrStreamSetPanes } return sw.worksheet.setPanes(panes) } // MergeCell provides a function to merge cells by a given range reference for // the StreamWriter. Don't create a merged cell that overlaps with another // existing merged cell. func (sw *StreamWriter) MergeCell(hCell, vCell string) error { _, err := cellRefsToCoordinates(hCell, vCell) if err != nil { return err } sw.mergeCellsCount++ _, _ = sw.mergeCells.WriteString(``) return nil } // setCellFormula provides a function to set formula of a cell. func setCellFormula(c *xlsxC, formula string) { if formula != "" { c.T, c.F = "str", &xlsxF{Content: formula} } } // setCellTime provides a function to set number of a cell with a time. func (sw *StreamWriter) setCellTime(c *xlsxC, val time.Time) error { var date1904, isNum bool wb, err := sw.file.workbookReader() if err != nil { return err } if wb != nil && wb.WorkbookPr != nil { date1904 = wb.WorkbookPr.Date1904 } if isNum, err = c.setCellTime(val, date1904); err == nil && isNum && c.S == 0 { style, _ := sw.file.NewStyle(&Style{NumFmt: 22}) c.S = style } return nil } // setCellValFunc provides a function to set value of a cell. func (sw *StreamWriter) setCellValFunc(c *xlsxC, val interface{}) error { var err error switch val := val.(type) { case int, int8, int16, int32, int64, uint, uint8, uint16, uint32, uint64: err = setCellIntFunc(c, val) case float32: c.T, c.V = setCellFloat(float64(val), -1, 32) case float64: c.T, c.V = setCellFloat(val, -1, 64) case string: c.setCellValue(val) case []byte: c.setCellValue(string(val)) case time.Duration: c.T, c.V = setCellDuration(val) case time.Time: err = sw.setCellTime(c, val) case bool: c.T, c.V = setCellBool(val) case nil: c.setCellValue("") case []RichTextRun: c.T, c.IS = "inlineStr", &xlsxSI{} c.IS.R, err = setRichText(val) default: c.setCellValue(fmt.Sprint(val)) } return err } // setCellIntFunc is a wrapper of SetCellInt. func setCellIntFunc(c *xlsxC, val interface{}) (err error) { switch val := val.(type) { case int: c.T, c.V = setCellInt(val) case int8: c.T, c.V = setCellInt(int(val)) case int16: c.T, c.V = setCellInt(int(val)) case int32: c.T, c.V = setCellInt(int(val)) case int64: c.T, c.V = setCellInt(int(val)) case uint: c.T, c.V = setCellInt(int(val)) case uint8: c.T, c.V = setCellInt(int(val)) case uint16: c.T, c.V = setCellInt(int(val)) case uint32: c.T, c.V = setCellInt(int(val)) case uint64: c.T, c.V = setCellInt(int(val)) default: } return } // writeCell constructs a cell XML and writes it to the buffer. func writeCell(buf *bufferedWriter, c xlsxC) { _, _ = buf.WriteString(``) if c.F != nil { _, _ = buf.WriteString(``) _ = xml.EscapeText(buf, []byte(c.F.Content)) _, _ = buf.WriteString(``) } if c.V != "" { _, _ = buf.WriteString(``) _ = xml.EscapeText(buf, []byte(c.V)) _, _ = buf.WriteString(``) } if c.IS != nil { if len(c.IS.R) > 0 { is, _ := xml.Marshal(c.IS.R) _, _ = buf.WriteString(``) _, _ = buf.Write(is) _, _ = buf.WriteString(``) } if c.IS.T != nil { _, _ = buf.WriteString(``) _, _ = buf.Write([]byte(c.IS.T.Val)) _, _ = buf.WriteString(``) } } _, _ = buf.WriteString(``) } // writeSheetData prepares the element preceding sheetData and writes the // sheetData XML start element to the buffer. func (sw *StreamWriter) writeSheetData() { if !sw.sheetWritten { bulkAppendFields(&sw.rawData, sw.worksheet, 4, 5) if sw.cols.Len() > 0 { _, _ = sw.rawData.WriteString("") _, _ = sw.rawData.WriteString(sw.cols.String()) _, _ = sw.rawData.WriteString("") } _, _ = sw.rawData.WriteString(``) sw.sheetWritten = true } } // Flush ending the streaming writing process. func (sw *StreamWriter) Flush() error { sw.writeSheetData() _, _ = sw.rawData.WriteString(``) bulkAppendFields(&sw.rawData, sw.worksheet, 8, 15) mergeCells := strings.Builder{} if sw.mergeCellsCount > 0 { _, _ = mergeCells.WriteString(``) _, _ = mergeCells.WriteString(sw.mergeCells.String()) _, _ = mergeCells.WriteString(``) } _, _ = sw.rawData.WriteString(mergeCells.String()) bulkAppendFields(&sw.rawData, sw.worksheet, 17, 38) _, _ = sw.rawData.WriteString(sw.tableParts) bulkAppendFields(&sw.rawData, sw.worksheet, 40, 40) _, _ = sw.rawData.WriteString(``) if err := sw.rawData.Flush(); err != nil { return err } sheetPath := sw.file.sheetMap[sw.Sheet] sw.file.Sheet.Delete(sheetPath) delete(sw.file.checked, sheetPath) sw.file.Pkg.Delete(sheetPath) return nil } // bulkAppendFields bulk-appends fields in a worksheet by specified field // names order range. func bulkAppendFields(w io.Writer, ws *xlsxWorksheet, from, to int) { s := reflect.ValueOf(ws).Elem() enc := xml.NewEncoder(w) for i := 0; i < s.NumField(); i++ { if from <= i && i <= to { _ = enc.Encode(s.Field(i).Interface()) } } } // bufferedWriter uses a temp file to store an extended buffer. Writes are // always made to an in-memory buffer, which will always succeed. The buffer // is written to the temp file with Sync, which may return an error. // Therefore, Sync should be periodically called and the error checked. type bufferedWriter struct { tmp *os.File buf bytes.Buffer } // Write to the in-memory buffer. The error is always nil. func (bw *bufferedWriter) Write(p []byte) (n int, err error) { return bw.buf.Write(p) } // WriteString write to the in-memory buffer. The error is always nil. func (bw *bufferedWriter) WriteString(p string) (n int, err error) { return bw.buf.WriteString(p) } // Reader provides read-access to the underlying buffer/file. func (bw *bufferedWriter) Reader() (io.Reader, error) { if bw.tmp == nil { return bytes.NewReader(bw.buf.Bytes()), nil } if err := bw.Flush(); err != nil { return nil, err } fi, err := bw.tmp.Stat() if err != nil { return nil, err } // os.File.ReadAt does not affect the cursor position and is safe to use here return io.NewSectionReader(bw.tmp, 0, fi.Size()), nil } // Sync will write the in-memory buffer to a temp file, if the in-memory // buffer has grown large enough. Any error will be returned. func (bw *bufferedWriter) Sync() (err error) { // Try to use local storage if bw.buf.Len() < StreamChunkSize { return nil } if bw.tmp == nil { bw.tmp, err = os.CreateTemp(os.TempDir(), "excelize-") if err != nil { // can not use local storage return nil } } return bw.Flush() } // Flush the entire in-memory buffer to the temp file, if a temp file is being // used. func (bw *bufferedWriter) Flush() error { if bw.tmp == nil { return nil } _, err := bw.buf.WriteTo(bw.tmp) if err != nil { return err } bw.buf.Reset() return nil } // Close the underlying temp file and reset the in-memory buffer. func (bw *bufferedWriter) Close() error { bw.buf.Reset() if bw.tmp == nil { return nil } defer os.Remove(bw.tmp.Name()) return bw.tmp.Close() }