What it is

Node.js bindings for libxl. Both Node 0.10 and Node 0.11 starting with 0.11.13 are supported. Iojs starting with version 1.0.3 is supported as well (older versions might work).

Compilation and Installation

Pull the library into your project with npm install libxl and require the module via

var xl = require('libxl');

LibXL installation

As this packages contains only bindings for the libxl library, the library itself is required for building and running the bindings.

Compilation Phase

Before the bindings are compiled, the install-libxl.js script pulls the latest version of the library from the XLware FTP server and unpacks it in deps/libxl. Therefore, no separate installation of libxl is necessary for building the bindings.

However, if you want to compile and run against a particular version of libxl, you can do so by manually unpacking the library archive into deps/libxl before building the bindings. This will bypass the install-libxl.js script and build the bindings against that specific version of the library.

Runtime

In order to load and use the bindings, the libxl library must be available in your dynamic library search path. This is achieved by either

Copying the library into your system library search path, e.g. /usr/lib on Linux.

Copying the library into the working directory where you run the script which uses the bindings. The name of the library file is libxl.so on Linux, libxl.dylib on Mac and libxl.dll on Windows.

Properly setting the LD_LIBRARY_PATH (Linux) or DYLD_LIBRARY_PATH (Mac) environment variable. For example, the following command will execute the demo.js script in the package directory without requiring libxl to be installed separately

LD_LIBRARY_PATH="`pwd`/deps/libxl/lib:`pwd`/deps/libxl/lib64:$LD_LIBRARY_PATH" node demo.js

Overriding the location of the compiled bindings

You can override the location where the Javascript wrapper looks for the libxl.node file by setting the NODE_LIBXL_PATH environment variable. This allows to distribute / deploy an application that uses the bindings to a system which runs on a different platform / architecture without recompiling the bindings there.

API

Usage

A new excel document is created via

var xlsBook = new xl.Book(xl.BOOK_TYPE_XLS);

or

var xlsxBook = new xl.Book(xl.BOOK_TYPE_XLSX);

(for xlsx documents). The document is written to disk via

xlsBook.writeSync('file.xls');

or

xlsBook.write('file.xls', callback);

and read back via

xlsBook.loadSync('file.xls');

or

xlsBook.load('file.xls', callback);

where callback will be called after the operation has completed, receiving an optional error object as argument if anything goes wrong.

IMPORTANT: See below for additional notes on the async implementation of libxl calls.

The Javascript API closely follows the C++ API described in the libxl documentation. For example, adding a new sheet and writing two cells works as

var sheet = xlsBook.addSheet('Sheet 1');
sheet.writeStr(1, 0, 'A string');
sheet.writeNum(1, 1, 42);

Functions whose C++ counterpart returns void or an error status have been implemented to return the respective instance, so it is possible to chain calls

sheet
    .writeStr(1, 0, 'A string');
    .writeNum(1, 1, 42);

Errors are handled by throwing exceptions.

Functions that return multiple values by reference in C++ (like Book::dateUnpack) return a object with the return values as properties.

See 'Differences...' below for a more detailed description of the methods whose behavior differs from their C++ counterpart.

IMPORTANT: The Javascript API enforces the types defined in its C++ counterpart for all function arguments; there is no implicit type casting. For example, passing a number to Sheet::writeStr will throw a TypeError instead of silently converting the string to a number.

Coverage

The bindings cover the current (version 3.5.4) libxl API completely.

Implementation details and differences w.r.t. the C++ API

Asynchroneous variants of libxl calls

The async variants of libxl calls implement the standard Node.js API for async functions: a callback is passed as last argument which is called once the operation has finished. The first argument of the callback is an error object, which is undefined if the operation completed without errors. Any results are passed as additional arguments to the callback.

IMPORTANT: While an async operation is pending, other operations (sync or async) on the same book object (and its descendants like sheets, formats and fonts) are not allowed and will throw an exception. However, multiple simultaneous operations on different books are allowed.

The following async functions are available:

• book.write / book.save, book.load are implemented asynchroneously. If you need synchroneous behavior you can use book.loadSync etc.
• book.writeRaw / book.saveRaw, book.loadRaw are implemented asynchroneously. book.saveRaw and its alias return the book data as second argument to the supplied callback. Use book.loadRawSync & friends for synchroneous behavior.
• book.addPicture has a async version book.addPictureAsync. The index of the new picture is passed as the second argument to the callback.
• book.getPicture has a async version book.getPictureAsync. Picture type and data are passed to the callback as second and third arguments.
• sheet.insertRow and sheet.insertCol are very slow and thus are also available as async implementations sheet.insertRowAsync and sheet.insertColAsync.

Interface differences

• book.write, book.writeRaw and their sync versions are also available as book.save etc.
• book.loadRaw / book.loadRawSync take a node buffer as argument
• book.writeRaw / book.writeRawSync return a node buffer
• book.getPicture returns an object with type and data properties. The data property is a node buffer containing the image data.
• book.addPicture and book.addPictureAsync are overloaded and can be called with either a file path or a node buffer, thus implementing both Book.AddPicture and Book.AddPicture2 from the libxl API.
• book.dateUnpack: Returns an object with year, month, date, hour, minute, seconds and mseconds properties.
• book.colorUnpack: Returns an object with red, green and blue properties.
• book.defaultFont: Returns an object with name and size properties.
• sheet.readStr & friends: If sheet.readXXX is provided with an object as optional second argument, the cell format is returned in the objects format property.
• sheet.getMerge: Returns an object with the rowFirst, rowLast, colFirst, colLast properties.
• sheet.getPrintFit: Returns either false or an object with the wPages and hPages properties.
• sheet.getNamedRange: Returns an object with rowFirst, rowLast, colFirst, colLast and hidden properties.
• sheet.namedRange: Returns an object with rowFirst, rowLast, colFirst, colLast, name, scopeId, and hidden properties.
• sheet.getTopLeftView: Returns an object with row and col properties.
• sheet.addrToRowCol: Returns an object with row, col, rowRelative, colRelative properties.

Other differences

• Book object creation: Books are not created via xlCreateBook and xlCreateXMLBook. Instead, object instances are directly constructed from the xl.Book constructor via either new xl.Book(xl.BOOK_TYPE_XLS) or new xl.Book(xl.BOOK_TYPE_XLSX)
• Accessing the parent book: sheet, format and font objects hold a reference to their parent book that can be accessed via the book property

Enum constants

All C enum constants provided by the library are available as constants on the library object, e.g. xl.NUMFORMAT_DATE or xl.PICTURETYPE_PNG.

Unlocking the API

If you have purchased a licence key from XLware, you can call book.setKey in order to unlock the library. As an alternative, you can build the key into the bindings by modifying api_key.h and rebuilding the library via node-gyp rebuild (you'll have to install node-gyp for this) or npm install in the package directory.

Platform support

The package supports Linux, Windows and Mac.

Tests

The bindings are fully covered with jasmine tests. If you have jasmine-node installed (via NPM), you can run the suite via

jasmine-node specs/

Reporting bugs

Please report any bugs or feature requests on the github issue tracker.

Roadmap

As the API is completely covered, I consider the bindings complete. New releases will only cover new libxl methods and fix bugs. If you identify parts of libxl that are particularily slow, asynchroneous version of those could be added as well. Note that only the latest version branch (0.2.x) is maintained and supported.

Credits

• Torben Fitschen wrote the install script which pulls the necessary libxl SDK before building.
• Martin Schröder for adding Mac support.
• Parts if this package were developed during slacktime provided by the awesome folks at Mayflower GmbH
• Alexander Makarenko wrote node-excel-libxl Though node-libxl is rewritten from scratch, this package served as the starting point.