Notes

Changes

  • In v2.28.4,
  • In v2.27.0, the output_callback now includes a url parameter.
  • In v2.25.2, updated the output_saveRows option to accept a function.
  • In v2.25.1, the output_callback can return modified data instead of a boolean.
  • In v2.22.4, added output_formatContent callback function which allows for extra formatting of cell content (e.g. convert '&''&').

Older Notes

  • In v2.22.2,
    • The data-attribute used by the output_dataAttrib can now be an empty string.
    • Update output_saveRows option, it now includes the ability to set a jQuery filter selector.
      • For example, if you want to only output rows with an active checkbox, set this option to '.checked' (default class added by the checkbox parser contained in the "parser-input-select.js" file).
      • To output only visible active checkboxes, set this option to '.checked:visible'.
      • To output rows without an active checkbox, set this option to ':not(.checked)', etc.
      • More details can be found in the options section.
    • Fix issue with output_hiddenColumns causing an empty output.
  • In v2.22.0,
    • Added the BOM back to the UTF-8 csv downloadable file to support unicode characters in Excel.
    • Add output_hiddenColumns which includes hidden columns in the output when true.
  • In v2.21.0, added the output_includeFooter option.
  • In v2.17.5, the need to modify the server's content disposition no longer exists.
  • In v2.17.0,
    • Added the output_ignoreColumns option & modified the output_callback parameters.
    • Added output_duplicateSpans option to duplicate (when true) colspan & rowspan content across cells.
  • In v2.16.4, added the output_encoding option.

Requirements

  • This widget will only work in tablesorter version 2.8+ and jQuery version 1.7+.
  • To download a file with the set filename and extension, the server's Content-disposition needs to be set
  • This widget now uses a method available in modern browsers (IE10+) to download files without the need for server side modifications
  • Support in older browsers (IE9 and older) have not been throughly tested.

Features & Usage

  • This widget can output the table data to:
    • JSON
    • Array of arrays
    • Text with any separator (comma, tab, etc).
  • Rowspan and colspan are supported by this widget. Please take note of additional information in the "Rowspan and colspan" section below.
  • Working with Excel:
    • To save a file for Excel
      • Set the separator to a comma (,).
      • Save the file with a csv extension, then load it into Excel.
      • Or, save the file with a xls extension; but when loading the file in Excel, you will likely see a message stating that the file is corrupt. Continue loading and you will see the expected data placed into each cell.
    • To copy the data into Excel
      • Set the separator as a tab (\t).
      • Open the data in a popup window, then copy the contents to the clipboard.
      • Paste from the clipboard directly into a cell in Excel.
  • This demo uses Bootstrap to create the download button dropdown for each table allowing the user to choose the settings
    • Using Bootstrap started out relatively difficult... any click inside would close the dropdown.
    • There are likely better alternatives as the download button and popup are all completely customizable.
    • The download dropdown code is overly complicated because it is set up for two tables & allows modifying almost all of the options.
    • The only thing needed to output the table code (download or open a popup) is to trigger the following event after all of your widget options are set as desired:
      $table.trigger('outputTable');

Known Issues

  • Some versions of Safari do not fully support direct downloading, nor will you be able to save the file with the set file name; for more details please refer to this issue.

Credit

Basic Setup

Using the default settings of this widget will:
  • Output csv to a popup window.
  • Only include the last header row (if there are more than one rows).
  • Only include filtered rows (all rows will be included, even if the filter widget isn't being used).
  • Will only output the table cell text (ignoring any HTML).
Of course, all of the above behavior can be modified using the options.
This is an example of how to setup a table using only the default settings:

HTML

<button class="download">Get CSV</button>
<table class="tablesorter">
    <!-- .... -->
</table>

Script

$(function () {
    var $table = $('table');

    $('.download').click(function(){
        // tell the output widget do it's thing
        $table.trigger('outputTable');
    });

    $table.tablesorter({
        theme: 'blue',
        widgets: ['zebra', 'output']
    });
});
Modification of the .htaccess is no longer required

.htaccess

The Content-disposition headers may need to be set in order to make downloads work properly in IE. So .htaccess can look like:
<FilesMatch "\.(?i:csv|txt)$">
	    ForceType application/octet-stream
	    Header set Content-Disposition attachment
	</FilesMatch>

Setup Example (php)

Thanks to TheSin- for sharing this setup which includes the necessary php to allow file download with the proper filename and extension (original Gist)

Rowspan and colspan

  • Even though tablesorter does not currently support rowspan and colspan within the tbody (except in child rows), this widget does support them.

  • When the widget encounters a colspan:
    • The colspan will be processed such that the data will be applied to each cell as if the colspan did not exist. So a header row that looks like this:
      <th>data</th><th colspan="2">values</th>
      will produce an output like this:
      data, values, values
    • Important But, if a JSON output is chosen, the resulting object for a row cannot contain duplicate keys.

      [{ "data" : "0", "values" : "1", "values" : "2" }]
      // becomes [{ "data" : "0", "values" : "2" }]
      So, in order to fix this, any key (header name) with a colspan will have the column index added to the end (see the output_callbackJSON option). So the same HTML above will produce the following output:
      [{ "data" : "0", "values" : "1", "values(2)" : "2" }]

  • When the widget encounters a rowspan:
    • The rowspan will be processed such that the data will be applied to each cell as if the rowspan did not exist. So if two header rows looks like this:
      <tr>
        <th rowspan="2">line</th>
        <th colspan="3">values</th>
      </tr>
      <tr>
        <th>value1</th>
        <th>value2</th>
        <th>value3</th>
      </tr>
      And the output_headerRows is true, the output will look like this:
      line,values,values,values
      line,value1,value2,value3
      1,1.1,1.2,1.3
      1,1.4,1.5,1.5
    • Important But, if a JSON output is chosen, and output_headerRows is set to false, only header names from the last row will be applied to the output. The HTML above will produce this output:

      [
        { "line" : "1", "value1" : "1.1", "value2" : "1.2", "value3" : "1.3" },
        { "line" : "1", "value1" : "1.4", "value2" : "1.5", "value3" : "1.5" }
      ]
      If the output_headerRows is set to true, the header names from the second row will be applied to alternating rows and the header with a colspan will get uniques names using the output_callbackJSON option. So the same HTML above will produce the following output:
      [
        { "line" : "1", "values" : "1.1", "values(2)" : "1.2", "values(3)" : "1.3" },
        { "line" : "1", "value1" : "1.4", "value2"    : "1.5", "value3"    : "1.5" }
      ]

Options

Output widget default options (added inside of tablesorter widgetOptions)

TIP! Click on the link in the option column to reveal full details (or toggle|show|hide all) or double click to update the browser location.

OptionDefaultDescription
',' This option allows you to set the output as JSON, Array, or change the separator between cell data with text output:

When the output is created, and it this option doesn't match either "json" or "array", each block of table cell data will be separated by this setting.

This is the result with this option set to "," and with the table filtered with Rank :
Rank,First,Last,Age,Total,Discount,Date
111,Peter,Parker,28,$9.99,20%,"Jul 6, 2006 8:14 AM"
166,Bruce Lee,Evans,22,$13.19,11%,"Jan 18, 2007 9:12 AM"
When this option is set to "array", this will be the resulting output:
[
    [ "Rank", "First", "Last", "Age", "Total", "Discount", "Date" ],
    [ "111", "Peter", "Parker", "28", "$9.99", "20%", "Jul 6, 2006 8:14 AM" ],
    [ "166", "Bruce Lee", "Evans", "22", "$13.19", "11%", "Jan 18, 2007 9:12 AM" ],
]
Note this requires JSON.stringify, or the result will be a flattened array.

And when this option is set to "json", the resulting output will be a valid JSON format:
[
    {
        "Rank" : "111",
        "First": "Peter",
        "Last": "Parker",
        "Age": "28",
        "Total": "$9.99",
        "Discount": "20%",
        "Date": "Jul 6, 2006 8:14 AM"
    }, {
        "Rank": "166",
        "First": "Bruce Lee",
        "Last": "Evans",
        "Age": "22",
        "Total": "$13.19",
        "Discount": "11%",
        "Date": "Jan 18, 2007 9:12 AM"
    }
]
Note this also requires JSON.stringify, or the result will show [ [object Object], ... ] in the output.
[ ] Add the zero-based column index to this array to ignore specific columns (v2.17.0)

For example, to use this option to hide the first column, set it as follows:
output_ignoreColumns : [ 0 ]

Note This option will work properly with tables that contain rowspans & colspans.
output_hiddenColumns false Include hidden columns (using display: none) in the output (v2.22.0).
output_includeFooter false Change this option to true to include the <tfoot> rows in the output (v2.21.0)
output_includeHeader true Change this option to false to not include any rows from the <thead> in the output (v2.28.4)
false Setting this option to true will include all header rows while processing a JSON output.

If this option is set to false, only the last row of the table header will be used as a key names for the tbody values; this does assume that the last table row in the header contains all of the header cell names & does not contain any colspan.
'data-name' This option allows you to override any table cell (thead or tbody) text with the contents of this data-attribute (v2.22.2)

In the below example, the header text is obtained from this set data-attribute. If the data-attribute doesn't exist, then the header name is obtained from the actual header text.

The data-attribute can be an empty string (v2.22.2).

<thead>
	<tr>
		<th>Rank</th>
		<th data-name="First">First Name</th>
		<th data-name="Last">Last Name</th>
		<th>Age</th>
		<th>Total</th>
		<th>Discount</th>
		<th>Date</th>
	</tr>
</thead>

Notice in this output that the "First" and "Last" columns match the contents of the data-name attribute and not the header text:

Rank,First,Last,Age,Total,Discount,Date
* NOTE * The core plugin uses the data-text attribute for text extraction. If you want to use the same data for both, make the data-attributes match!
'popup' Change this option to either 'popup' or 'download' to set the destination of the output.

When setting this option, only the first letter is required ('p' or 'd') :
output_delivery : 'p'
'filtered' Change this option to either 'filtered', 'visible', 'all', a jQuery filter selector (v2.25.1) to set to match rows to be added to the output, or a jQuery filter callback function (v2.25.2).

  • 'filtered' - output rows that have not been filtered by the filter widget (i.e. rows that do not have the "filtered" class applied; this class name is set by the filter_filteredRow widget option).
  • 'visible' - only output rows that are currently visible; rows hidden by pagination will not be included.
  • jQuery filter selector - output all rows that match the jQuery filter (use a string; not a jQuery object).
    • For example, '.checked', '.checked:visible' or :not(.checked) can be used; these examples match the class set by the checkbox parser table.config.checkboxClass class setting.
    • For reference, internally, the rows are filtered using this setting: $rows.filter( wo.output_saveRows ).
    • Instead of checking to see if output_saveRows works as a filter, the option is checked to see if it starts with any of the following characters before it is used as a filter:
      • '.' (class selector)
      • '#' (id selector)
      • ':' (filter selectors, e.g. ':visible' or ':not()')
      • [ (start of an attribute selector)
  • 'all' - output all rows even if they are hidden. When the pager removeRows option is true, rows not visible do not exist in the DOM and therefore will not be included.
  • function - (v2.25.2) Include a filter function to test each row and return a boolean value; true to include the row, and false to exclude. For example:
    output_saveRows: function(index, element) {
      // "this" is the same as "element"
      // only include the row if the select element's value is set to "include".
      return $(this).find('select').val() === 'include';
    }

Even if this option is set to 'filtered' and the filter widget is not being used, all of the rows will be added to the output.

*NOTE* When setting this option,

  • Only the first letter is required for filtered ('f') or visible ('v') rows.
    output_saveRows : 'f'
  • If using a jQuery selector, it won't be recognized as a selector unless one of the characters (.#:[) is at the start of the string; if your filter is not working, or uses a charater not listed, then use the callback function, or you can contact me for support.
  • If no matches are found, all rows will be sent to the output.
true When true, colspan & rowspan content is duplicated in the output

By default, any tbody cells that are included in the colspan or rowspan will have the cell's content duplicated in the output. When set to false, the cells within the colspan or rowspan will be empty.
Here is an example of the second table output with this option set to false:
line,values,values,values
line,value1,value2,value3
1,1.1,1.2,1.3
,1.4,1.5,
2,2.1,2.2,2.3
,2.4,2.5,
3,3.1,3.2,3.3
,3.4,3.5,
4,4.1,,4.2
,,,4.3
This option does not affect thead cells, they will always have duplicated content.
'\u201c;' When a table cell contains a quote, it needs to be modified to differentiate it from cell content that is wrapped in quotes.

Quotes are necessary to maintain the integrity of the output data when any cell content contains either the separator or carriage returns. In order to do this, this widget replaces any quotes within cell content with this character.

By default, this replacement character is a left double quote (, shown as it's unicode equivalent '\u201c;').
output_replaceQuote : '\u201c;' // left double quote
In the demo below, additional examples of a single quote (') and escaped double quote (\") are added as selectable buttons.
Rank,First,Last,Age,Total,Discount,Date
1,Philip Aaron Wong,Johnson Sr Esq,25,$5.95,22%,"Jun 26, 2004 7:22 AM"
21,John,“Robin“\n\tHood,33,$19.99,25%,"Dec 10, 2002 5:14 AM"
013,Clark,“Old man“ Kent Sr.,18,$15.89,44%,"Jan 12, 2003 11:14 AM"
16,Jim,“Jimmy“ Franco,24,-$14.19,24%,"Jan 14, 2004 11:23 AM"
55,Dennis,“Charley“ Bronson,65,$123.00,32%,"Jan 20, 2001 1:12 PM"
false This option, if true tells the widget to include any HTML within the table cells.

Only cells within the table body will include HTML. The table header has a lot of extra markup for the sorting arrows, so it is automatically stripped of HTML.
true This option, if true tells the widget to remove any leading and trailing (white)space characters

Leading and trailing whitespace characters include newlines, spaces, non-breaking spaces and tabs. Any of these whitespace characters in the middle of the table cell will always be preserved.

*NOTE* carriage returns in the middle of the table cell content will be stripped out if this option is set to true.
false This option, if true wrap the output of all table cell content in quotes.

This is the output of the widget when the age column is filtered for results in the range .
"Rank","First","Last","Age","Total","Discount","Date"
"1","Philip Aaron Wong","Johnson Sr Esq","25","$5.95","22%","Jun 26, 2004 7:22 AM"
"111","Peter","Parker","28","$9.99","20%","Jul 6, 2006 8:14 AM"
"21","John","'Robin'\n\tHood","33","$19.99","25%","Dec 10, 2002 5:14 AM"
"9","Martha","delFuego","25","$22.09","12%","Jun 11, 2011 10:55 AM"
Notice that every block of data is wrapped in a quote. When this option is false, only blocks that contain a newline or separator (comma, by default) will be wrapped in quotes. Here is the same data set without this option set:
Rank,First,Last,Age,Total,Discount,Date
1,Philip Aaron Wong,Johnson Sr Esq,25,$5.95,22%,"Jun 26, 2004 7:22 AM"
111,Peter,Parker,28,$9.99,20%,"Jul 6, 2006 8:14 AM"
21,John,'Robin'\n\tHood,33,$19.99,25%,"Dec 10, 2002 5:14 AM"
9,Martha,delFuego,25,$22.09,12%,"Jun 11, 2011 10:55 AM"
Only the dates, which contain commas, are wrapped in quotes.
'width=500,height=300' If the data is sent to a popup window, the features set by this option are applied (do not include any spaces).

Set the desired popup window features within this string. For a full list of features, please refer to the Mozilla Developer Network: Window.open documentation page (look for strWindowFeatures parameters).

The <textarea> within the popup window is set with 100% width and height.
output_saveFileName 'mytable.csv' When downloading, this option sets the filename of the output.
null This callback is executed after the content of a table cell has been processed.

Default setting & how to use this option:
output_formatContent : function( c, wo, data ) {
	// data.isHeader (boolean) = true if processing a header cell
	// data.$cell = jQuery object of the cell currently being processed
	// data.content = processed cell content (spaces trimmed, quotes added/replaced, etc)
	// **********
	// use data.$cell.html() to get the original cell content
	return data.content
		.replace(/fred/ig, '')  // because we don't like fred
		// because HTML is being processed, you might want to also replace HTML codes
		.replace(/&amp;/g, '&') // convert '&amp;' → '&'
}
If you want to replace all HTML codes, check out the he library; then you can use this code (Fred would be happy):
output_formatContent : function( c, wo, data ) {
	// replace all HTML shortcut codes
	// (e.g. 'foo &copy; bar &ne; baz &#x1D306; qux' → 'foo © bar ≠ baz 𝌆 qux' )
	return he.decode( data.content );
}
{see description} Return true or modified data in this callback to continue the download or open the popup (v2.25.1).

In v2.27.0, a url parameter was added to the callback.

In v2.25.1, this callback can return modified data for output.

As of v2.17.0 (v2.27.0) , the callback function was modified to pass two three parameters:

  • config - the table.config settings.
  • data - the output data as a string
  • url - the url of the current page (including filters, sort, etc), if the pager widget/addon is being used.
    • You can get the same value by using config.pager.ajaxObject.url; this parameter was added for convenience.
    • The value will be null if the pager is not being used.
Default setting:
function(config, data, url){
    // return data.replace( /Mary Smith/g, 'Mary Smith-Jones' ); // v2.25.1+
    return true;
}
After table processing has completed, this callback function is exectued. If true is not returned, the processed data will not open a popup, nor download the data.
output_callback : function(config, data, url) {
    // send output to the console only
    console.log(data);
    return false;
}
If you are using the pager and want to provide the entire table in the output, use this callback as follows:
function(config, data, url) {
    // This is only an example, but you could set it up so that a "&csv=1"
    // parameter in the url to signal the server to provide the csv for the
    // entire table using the current filters, sort, etc.
    if (url) {
      return $.getJSON(url + "&csv=1")
        .done(function(data){
          return data.csv;
        })
        .fail(function() {
          alert('no go on CSV');
          return false;
        });
    }
    return true;
}
{see description} JSON callback executed when a colspan is encountered in the header.

Default setting & how to use this option:
output_callbackJSON : function($cell, txt, cellIndex) {
    return txt + '(' + (cellIndex + col) + ')';
}
For more details on why this option is necessary, see the "Rowspan and colspan" section above.
{see description} Default encoding setting (2.16.4; 2.17.5)

As of 2.17.5, this option no longer needs to be modified to specifically make this widget download files that will work in Excel.

The method used to download has been completely changed. The downloads still need an encoding setting, but this option is now set to a default of
output_encoding : 'data:application/octet-stream;charset=utf8,'

The information below is no longer relavant:
With the default settings (utf-8 no BOM), Excel does not properly encode accented characters unless the csv file is imported. Depending on the characters used, there are various methods which will allow proper encoding, but no one method is ideal. So this option can be set to allow the user to try different encodings. Set it as follows:
// output data type (with BOM or Windows-1252 is needed for excel)
// NO BOM   : 'data:text/csv;charset=utf8,'
// With BOM : 'data:text/csv;charset=utf8,%EF%BB%BF'
// WIN 1252 : 'data:text/csv;charset=windows-1252,' // ANSI (subset of ISO-8859-1)
output_encoding : 'data:text/csv;charset=utf8,'
Note The commas are important!
null Override internal save file code and use an external plugin (v2.28.4)

You can use a plugin such as FileSaver.js instead of the simplified, and possibly out-of-date internal method.

Use this option as follows:
output_savePlugin: function(config, widgetOptions, data) {
  var blob = new Blob([data], {type: widgetOptions.output_encoding});
  saveAs(blob, widgetOptions.output_saveFileName);
}
Note If you need to change the carriage return and/or the tab replacement strings, modify them as follows (changed in v2.21.2):
$.tablesorter.output.replaceCR = '\x0d\x0a';
$.tablesorter.output.replaceTab = '\x09';

Demo


Search in the Discount column

Table showing filter & pager support

  Rank First Name Last Name Age Total Discount Date
Rank First Name Last Name Age Total Discount Date
1Philip Aaron WongJohnson Sr Esq25$5.9522%Jun 26, 2004 7:22 AM
11Aaron"doc" Hibert12$2.995%Aug 21, 2009 12:21 PM
12Brandon ClarkHenry Jr51$42.2918%Oct 13, 2000 1:15 PM
111PeterParker28$9.9920%Jul 6, 2006 8:14 AM
21John "Robin"
Hood
33$19.9925%Dec 10, 2002 5:14 AM
013Clark"Old man" Kent Sr.18$15.8944%Jan 12, 2003 11:14 AM
005BruceAlmighty Esq45$153.1944%Jan 18, 2021 9:12 AM
10AlexDumass13$5.294%Jan 8, 2012 5:11 PM
16Jim"Jimmy" Franco24-$14.1924%Jan 14, 2004 11:23 AM
166Bruce LeeEvans22$13.1911%Jan 18, 2007 9:12 AM
100Brenda DexterMcMasters18$55.2015%Feb 12, 2010 7:23 PM
55Dennis"Charley" Bronson65$123.0032%Jan 20, 2001 1:12 PM
9MarthadelFuego25$22.0912%Jun 11, 2011 10:55 AM

Table showing output widget rowspan & colspan support

line values
value1 value2 value3
1 1.1 1.2 1.3
1.4 1.5
2 2.1 2.2 2.3
2.4 2.5
3 3.1 3.2 3.3
3.4 3.5
4 4.1 4.2
4.3

Page Header

<!-- Tablesorter: required -->
<link href="css/theme.bootstrap.css" rel="stylesheet">
<script src="js/jquery-latest.min.js"></script>
<script src="js/jquery.tablesorter.js"></script>
<script src="js/jquery.tablesorter.widgets.js"></script>
<!-- pager -->
<link href="css/jquery.tablesorter.pager.css" rel="stylesheet">
<script src="js/jquery.tablesorter.pager.js"></script>
<!-- Ouput widget -->
<script src="js/parsers/parser-input-select.js"></script>
<script src="js/widgets/widget-ouput.js"></script>

CSS


	

Script


	

HTML