Fetching contributors…
Cannot retrieve contributors at this time
473 lines (346 sloc) 14 KB


Node-XLSX-Stream is written in literate coffeescript. The following is the actual source of the module.

fs = require('fs')
blobs = require('./blobs')
Archiver = require('archiver')

module.exports = class XlsxWriter

Simple writes

XlsxWriter.write(out: String, data: Array, cb: Function)

The simplest way to use Node-XLSX-Stream is to use the write method.

The callback comes directly from fs.writeFile and has the arity (err)

  # @param {String} out Output file path.
  # @param {Array} data Data to write.
  # @param {Function} cb Callback to call when done. Fed (err).
  @write = (out, data, cb) ->
    writer = new XlsxWriter({out: out})

Advanced usage

Node-XLSX-Stream has more advanced features available for better customization of spreadsheets.

When constructing a writer, pass it an optional file path and customization options.

new XlsxWriter([options]: Object) : XlsxWriter
  # Build a writer object.
  # @param {Object} [options] Preparation options.
  # @param {String} [options.out] Output file path.
  # @param {Array}  [options.columns] Column definition. Must be added in constructor.
  # @example options.columns = [
  #     {  width: 30 }, // width is in 'characters'
  #     {  width: 10 }
  # ]
  constructor: (options = {}) ->

    # Support just passing a string path into the constructor.
    if (typeof options == 'string')
      options = {out: options}

    # Set options.
    defaults = {
      defaultWidth: 15
      zip: {
        forceUTC: true # this is required, zips will be unreadable without it
      columns: []
    @options = _extend(defaults, options)

    # Start sheet.

    # Write column definition.

    # Create Zip.
    zipOptions = || {}
    zipOptions.forceUTC = true # force this on in all cases for now, otherwise we're useless
    @zip = Archiver('zip', zipOptions)

    # Archiver attaches an exit listener on the process, we don't want this,
    # it will fire if this object is never finalized.
    @zip.catchEarlyExitAttached = true

    # Hook this passthrough into the zip stream.
    @zip.append(@sheetStream, {name: 'xl/worksheets/sheet1.xml'})

Adding rows

Rows are easy to add one by one or all at once. Data types within the sheet will be inferred from the data types passed to addRow().

addRow(row: Object)

Add a single row.

  # @example (javascript)
  # writer.addRow({
  #     "A String Column" : "A String Value",
  #     "A Number Column" : 12345,
  #     "A Date Column" : new Date(1999,11,31)
  # })
  addRow: (row) ->

    # Values in header are defined by the keys of the object we've passed in.
    # They need to be written the first time they're passed in.
    if !@haveHeader
      col = 1
      for key of row
        @_addCell(key, col)
        col += 1

      @haveHeader = true

    for key, col in @cellMap
      @_addCell(row[key] || "", col + 1)
addRows(rows: Array)

Rows can be added in batch.

  addRows: (rows) ->
    for row in rows
defineColumns(columns: Array)

Column definitions can be easily added, but it must be done before rows are added to prevent a nasty Excel bug.

  # @example (javascript)
  # writer.defineColumns([
  #     {  width: 30 }, // width is in 'characters'
  #     {  width: 10 }
  # ])
  defineColumns: (columns) ->
    if @haveHeader
      throw new Error """
        Columns cannot be added after rows! Unfortunately Excel will crash
        if column definitions come after sheet data. Please move your `defineColumns()`
        call before any `addRow()` calls, or define options.columns in the XlsxWriter
    @options.columns = columns
    # Write column metadata.
    # Would really like to do this at the end so that we don't have to mandate
    # it comes first, but Excel pukes if <cols> comes before <sheetData>.

File generation

Once you are done adding rows & defining columns, you have a few options for generating the file. The writeToFile helper is a one-stop-shop for writing directly to a file using fs.writeFile; otherwise, you can pack() manually, which will return a readable stream.

writeToFile([fileName]: String, cb: Function)

Writes data to a file. Convenience method.

If no filename is specified, will attempt to use the one specified in the constructor as options.out.

The callback is fed directly to fs.writeFile.

  # @param {String} [fileName] File path to write.
  # @param {Function} cb Callback.
  writeToFile: (fileName, cb) ->
    if fileName instanceof Function
      cb = fileName
      fileName = @options.out
    if !fileName
      throw new Error("Filename required. Supply one in writeToFile() or in options.out.")

    # Create zip, pipe it into a file writeStream.
    zip = @createReadStream(fileName)
    fileStream = fs.createWriteStream(fileName)
    fileStream.once 'finish', cb
createReadStream() : Stream Deprecated
  # @return {Stream} Readable stream with ZIP data.
  createReadStream: () ->
getReadStream() : Stream

Returns a readable stream from this file. You can pipe this directly to a file or response object. Be sure to use 'binary' mode.

You are responsible for indicating that you have finished the file generation by calling finalize(), which will end the sheet stream.

  # @return {Stream} Readable stream with ZIP data.
  getReadStream: () ->

Finishes up the sheet & generate shared strings. You must call this manually if you are using createReadStream.

  finalize: () ->

    if @finalized
      throw new Error "This XLSX was already finalized."

    # Mark this as finished
    @finalized = true

    # If there was data, end sheetData
    if @haveHeader

    # Write relationships data.

    # Generate shared strings

    # Generate external rels

    # End sheet

    # Add supporting files to zip and finalize it. The readStream (@zip) will soon emit
    # an 'end' event.

Cancel use of this writer and close all streams. This is not needed if you've written to a file.

  dispose: () ->
    return if @disposed

    while( # drain stream
      1; # noop
    delete @zip
    delete @sheetStream
    @disposed = true

Internal methods

Adds a cell to the row in progress.

  # @param {String|Number|Date} value Value to write.
  # @param {Number}             col   Column index.
  _addCell: (value = '', col) ->
    row = @currentRow
    cell = @_getCellIdentifier(row, col)

    # Hyperlink support
    if == '[object Object]'
      if !value.value || !value.hyperlink
        throw new Error("A hyperlink cell must have both 'value' and 'hyperlink' keys.")
      @_addCell(value.value, col)
      @_createRelationship(cell, value.hyperlink)

    if typeof value == 'number'
      @rowBuffer += blobs.numberCell(value, cell)
    else if value instanceof Date
      date = @_dateToOADate(value)
      @rowBuffer += blobs.dateCell(date, cell)
      index = @_lookupString(value)
      @rowBuffer += blobs.cell(index, cell)

Begins a row. Call this before starting any row. Will start a buffer for all proceeding cells, until @_endRow is called.

  _startRow: () ->
    @rowBuffer = blobs.startRow(@currentRow)
    @currentRow += 1

Ends a row. Will write the row to the sheet.

  _endRow: () ->
    @_write(@rowBuffer + blobs.endRow)

Given row and column indices, returns a cell identifier, e.g. "E20"

  # @param {Number} row  Row index.
  # @param {Number} cell Cell index.
  # @return {String}     Cell identifier.
  _getCellIdentifier: (row, col) ->
    colIndex = ''
    if @cellLabelMap[col]
      colIndex = @cellLabelMap[col]
      if col == 0
        # Provide a fallback for empty spreadsheets
        row = 1
        col = 1

      input = (+col - 1).toString(26)
      while input.length
        a = input.charCodeAt(input.length - 1)
        colIndex = String.fromCharCode(a + if a >= 48 and a <= 57 then 17 else -22) + colIndex
        input = if input.length > 1 then (parseInt(input.substr(0, input.length - 1), 26) - 1).toString(26) else ""
      @cellLabelMap[col] = colIndex

    return colIndex + row

Creates column definitions, if any definitions exist. This will write column styles, widths, etc.

  # @return {String} Column definition.
  _generateColumnDefinition: () ->
    # <cols/> tag (empty) crashes excel, weeeeee
    if !@options.columns || !@options.columns.length
      return ''

    columnDefinition = ''
    columnDefinition += blobs.startColumns

    idx = 1
    for index, column of @options.columns
      columnDefinition += blobs.column(column.width || @options.defaultWidth, idx)
      idx += 1

    columnDefinition += blobs.endColumns
    return columnDefinition

Generates StringMap XML. Used as a finalization step - don't call this while building the xlsx is in progress.

Saves string data to this object so it can be written to the zip.

  _generateStrings: () ->
    stringTable = ''
    for string in @strings
      stringTable += blobs.string(@escapeXml(string))
    @stringsData = blobs.stringsHeader(@strings.length) + stringTable + blobs.stringsFooter

Looks up a string inside the internal string map. If it doesn't exist, it will be added to the map.

  # @param {String} value String to look up.
  # @return {Number}      Index within the string map where this string is located.
  _lookupString: (value) ->
    if !@stringMap[value]
      @stringMap[value] = @stringIndex
      @stringIndex += 1
    return @stringMap[value]

Create a relationship. For now, this is always a hyperlink. This writes to a array that will later be used define the rels.

  _createRelationship: (cell, target) ->
    @relationships.push({cell: cell, target: target})

Generate external relationships data. This is saved into "xl/worksheets/_rels/sheet1.xml.rels".

  _generateRelationships: () ->
    @relsData = blobs.externalWorksheetRels(@relationships)

Converts a Date to an OADate. See this stackoverflow post

  # @param {Date} date Date to convert.
  # @return {Number}   OADate.
  _dateToOADate: (date) ->
    epoch = new Date(1899,11,30)
    msPerDay = 8.64e7

    v = -1 * (epoch - date) / msPerDay

    # Deal with dates prior to 1899-12-30 00:00:00
    dec = v - Math.floor(v)

    if v < 0 and dec
      v = Math.floor(v) - dec

    return v

Convert an OADate to a Date.

  # @param {Number} oaDate OADate.
  # @return {Date}         Converted date.
  _OADateToDate: (oaDate) ->
    epoch = new Date(1899,11,30)
    msPerDay = 8.64e7

    # Deal with -ve values
    dec = oaDate - Math.floor(oaDate)

    if oaDate < 0 and dec
      oaDate = Math.floor(oaDate) - dec

    return new Date(oaDate * msPerDay + +epoch)

Resets sheet data. Called on initialization.

  _resetSheet: () ->

    # Sheet data storage.
    @sheetData = ''
    @strings = []
    @stringMap = {}
    @stringIndex = 0
    @stringData = null
    @currentRow = 0

    # Cell data storage
    @cellMap = []
    @cellLabelMap = {}

    # Column data storage
    @columns = []

    # Rels data storage
    @relData = ''
    @relationships = []

    # Flags
    @haveHeader = false
    @finalized = false

    # Create sheet stream.
    PassThrough = require('stream').PassThrough
    @sheetStream = new PassThrough()

    # Start off the sheet.

Finalizes this file and adds supporting docs. Should not be called directly.

  _finalizeZip: () ->
      .append(blobs.contentTypes, {name: '[Content_Types].xml'})
      .append(blobs.rels, {name: '_rels/.rels'})
      .append(blobs.workbook, {name: 'xl/workbook.xml'})
      .append(blobs.styles, {name: 'xl/styles.xml'})
      .append(blobs.workbookRels, {name: 'xl/_rels/workbook.xml.rels'})
      .append(@relsData, {name: 'xl/worksheets/_rels/sheet1.xml.rels'})
      .append(@stringsData, {name: 'xl/sharedStrings.xml'})

Wrapper around writing sheet data.

  # @param {String} data Data to write to the sheet.
  _write: (data) ->

Utility method for escaping XML - used within blobs and can be used manually.

  # @param {String} str String to escape.
  escapeXml: (str = '') ->
    return str.replace(/&/g, '&amp;')
      .replace(/</g, '&lt;')
      .replace(/>/g, '&gt;')

Simple extend helper.

_extend = (dest, src) ->
  for key, val of src
    dest[key] = val