OpenOffice::OODoc::Text - The text processing submodule of OpenOffice::OODoc


This man chapter describes the text-oriented methods of OpenOffice::OODoc, implemented by the OpenOffice::OODoc::Text class, and inherited by the OpenOffice::OODoc::Document class.

This class should not be explictly used in an ordinary application, because all its features are available in the OpenOffice::OODoc::Document class, in combination with other features. Practically, the present manual is provided to describe the text-oriented features of OpenOffice::OODoc::Document (knowing that these features are technically supported by the OpenOffice::OODoc::Text component of the API).

These features are text-oriented, but can be used against any class document, such as spreadsheets or presentations as well as text documents.

The OpenOffice::OODoc::Text class is a specialist derivative of OpenOffice::OODoc::XPath for XML elements which describe the text content of documents. Here, "text content" means containers that can host text content (i.e. tables, lists...) as well as flat text.

Knowing that the "styles.xml" member of an file can contain text (because some style definitions, such as page headers or footers, can contain text), the presently described features can be used against this member as well as the "content.xml" member.

This module Should be used with OpenOffice::OODoc::Styles, via the OpenOffice::OODoc::Document class, if the application has to handle detailed presentation parameters of text elements. This is because such parameters are held in styles elements and not in the text elements themselves, according to the principle of separation of content and presentation which is one of the foundations of the format.


Constructor : OpenOffice::OODoc::Text->new(<parameters>)

        Short Form: ooText(<parameters>)

        See OpenOffice::OODoc::XPath->new

        The XML member loaded by default is 'content.xml'. The most common
        creation method is like this:

            my $doc = OpenOffice::OODoc::Text->new
                (file => 'my_file.sxw');

        Other parameters can be supplied as options (see the properties list
        at the end of the chapter).


            my %delim =
                'text:h'                =>
                        begin   => '\sect{',
                        end     => '}'
                'text:list-item'        =>
                        begin   => '\item'
                'text:footnote-body' =>
                        begin   => '\footnote{',
                        end     => '}'
            my $doc = OpenOffice::OODoc::Text->new
                        file            => 'filename.sxw',
                        paragraph_style => 'My Paragraphs',
                        header_style    => 'My Headers',
                        delimiters      => { %delim }

        This technique gives the default styles to be used when creating new
        text elements. It also gives the particular delimiters (in this case
        LaTeX style markers) to be used at the beginning or end of some
        elements (in this case headers, list elements, footers) where the
        text is to be exported "as is". See the getText method of
        OODoc::Text for information about exporting text.

appendBodyElement(element [, options])

        Copies an existing element of any type and appends it to the end of
        the document body. No new element is created.


        Creates a new header of any level and appends it to the end of the

        Options are given as a hash [key => value]:

            'text'              => <header text>
            'level'             => header level, default is 1
            'style'             => header style, default is 'Heading 1'


            $doc->appendHeader(text => 'Next section');

        adds the text 'Next section' as 'Header 1'.

                text    => 'Chapter Conclusion',
                level   => '2',
                style   => 'Heading 2'

        adds a level 2 header to the end of the text body.

        You can give any XML attribute to the new header except for style or
        header level. In this case, the program must construct a hash
        containing pairs of key-values for the attributes you want to create
        and pass it using the 'attribute' option. Example:

            my %attr    = ( 'att1' => 'value1', 'att2' => 'value2' );
                text    => 'Attributes are important',
                level   => '1',
                style   => 'Chapter header',
                attribute => %attr

        If the 'text' option is empty, the header is created with an empty

        Note: this method can only be used with a new header i.e. it adds
        while it creates. To add an already available element using
        getHeader from the same document or from another document, use the
        appendElement method instead which is inherited from OODoc::XPath.

appendItem(list, text => text [,style => style ,[other_options]])

        Adds a new item to a list (ordered or unordered).

        The first argument is the existing list element (created using
        getOrderedList or getUnorderedList, for example). Options are the
        same as for appendParagraph.

        If the 'style' option is absent, the element is inserted according
        to the following rule:

appendItemList([type => list_type, [style => style [, options]]])

        Creates a new (empty) list and appends it to the end of the

        In 1 documents, an unordered list is the default,
        and if the 'type' option is given with the value 'ordered', then an
        ordered list is created. In Open Documents, the 'type' option is
        ignored because there are generic lists only.

        The 'style' options controls the list's style (as opposed to each
        item's style). If absent, the list takes the default paragraph style
        (see appendParagraph).

        Like appendParagraph, this method actually creates a new list
        element. To copy an existing list in the same document or in
        another, use appendElement or replicateElement instead.


        Creates a new paragraph and appends it to the end of the document.


            'text'              => <paragraph text>

            'style'             => <paragraph style>

        An 'attribute' option is also available under the same conditions as
        for the appendHeader method (see above) [19] .

        If the 'text' option is empty, calling this method is the equivalent
        of adding a line feed.

        If the 'style' option is empty, the style from the 'paragraph_style'
        property of the OODoc::Text instance is used.

        Note: this method can only be used with a new paragraph i.e. it adds
        while it creates. To add an already existing paragraph using
        getParagraph from the same document or from another document, use
        the appendElement or replicateElement methods instead which are
        inherited from OODoc::XPath, or even appendText below.

appendRow(table [, options])

        Appends a row to the end of the given table either by reference, by
        logical name or by sequential number. By default, the new row is
        simply an exact copy of the preceding row (in terms of content and
        presentation). You can pass an options hash which will give certain
        attributes to the created row, under the same conditions as for the
        appendElement method of OODoc::XPath. The returned value is the
        created row element.


            open SRC, '<', 'data.txt';
            my $table = $doc->getTable("Table1");
            my ($h, $l) = $doc->getTableSize($table);
            for (my $i = 0 ; my $record = <SRC> ; $i++)
                last unless $record;
                chomp $record;
                my @data = split ';', $record;
                my $row = $i < $h ?
                        $doc->getRow($table, $i) :
                for (my $j = 0 ; $j < $l ; $j++)
                        $doc->cellValue($row, $j, $data[$j]);

        The above program reads a CSV format data file sequentially (one
        record per line, comma-separated fields). Each record is split and
        put into a row in table Table1. On reading each new record, the
        reference for the following row is loaded by getRow, until the total
        number of rows is reached (total obtained previously using
        getTableSize). If the table is already full, it is lengthened by a
        row using appendRow. The internal loop loads the read data into the
        row's cells (pre-existing or newly created). See the sections on
        getTable, getRow, getTableSize and cellValue for a better
        understanding of this example.

        However, if good performance is what you are after, massive
        repetition of this method is not recommended (e.g. for lengthening a
        table dynamically, row by row, whilst loading external data into
        it). Rather than running dozens or hundreds of successive
        appendRows, it would be better for the application to read the total
        number of records to be loaded (using, for example, select count if
        from a relational database or otherwise preloading the data into an
        ordinary Perl table) and create a table of appropriate size in
        advance using insertTable or appendTable.

appendTable(name, rows, columns [, options])

        Creates a new table with the given name, number of rows and number
        of columns, and appends it by default to the end of the document
        body. The name must be unique within the document (the call is
        rejected if the name already exists). Returns the created table
        element if successful.

        'rows' and/or 'columns', if omitted, are replaced by the 'max_rows'
        and 'max_cols' properties of the document (see the properties below).

        By default, the table is set to fit the entire width between the
        left and right margins with equal sized columns, cells of type
        string and without borders or background colour.

        Possible options:

            'table-style'       => table style
            'cell-type'         => default cell type
            'cell-style'        => default cell style
            'text-style'        => default cell text style

        The first option is the name of a table style [20]  which defines
        certain global properties for the table (width, background colour,

        The second option is the cells' default data type. The main types
        available are string, float, currency, date, percentage. Caution: to
        be properly treated as having a numeric format in, a
        cell needs more than to be just marked 'numeric'. If the cell really
        needs to be treated properly as a number, you must also give it a
        cell style which itself refers to a number style. The cell-style
        parameter can do this. However, even though the OODoc::Styles module
        is there to otherwise help you create and add styles from a program,
        this type of exercise can become very labour-intensive. We therefore
        recommend using basic tables created in advance from document
        templates or style libraries created from an office application,
        rather than creating complex number tables from code.

        The text-style option selects the paragraph style applicable to the
        text displayed in each cell.

        Once the table is created, you can obviously modify each cell's type
        and style individually.


            my $table = $doc->appendTable
                                "Rate", 22, 5,
                                'table-style' => 'Table1',
                                'text-style' => 'Text body'


        See appendRow.

appendText(element_name [, <options>])

appendText(element [, <options])

        Appends a text element, by default to the end of a document.

        Two type of usage are possible:

        If the 'attachment' option is used, which indicates an element
        reference, the new element is attached as a "child" element of the
        given element. This allows you to place text in special zones, other
        than in the document body, which is not appropriate in all

        This method should be used to append unusual text elements (i.e. not
        paragraphs or headers) or existing elements (in its second usage
        type) of any type.

        Remember that:

            $doc->appendText('text:p', text => 'My text'); [21]

        is the same as:

            $doc->appendParagraph(text => 'My text');

bibliographyEntryContent(id [, key1 => value1, key2 => value2, ...])

        Gets, and optionnally sets, the properties of a given (existing)
        bibliographic entry. The optionally updated properties are provides
        as a hash. The returned description is a hash.

        The first argument can be either the logical identifier of the entry
        (as it appears for the end-user) or a previously found bibliography
        entry element (see getBibliographyElements()).


                my %desc = $doc->bibliographyEntryContent
                                        author  => 'Genicorp',
                                        pages   => 62
        This sequence updates the "Author" and "Pages" values of the "GEN99"
        entry, then returns all the content of the entry in %desc.

        Caution: Several bibliography entries can have the same identifier.
        This method processes one element at a time. In the example above,
        only the first occurrence of the "GEN99" entries is updated. So, if
        the user needs to ensure that all the entries with the same identifier
        have the same content, the appropriate code should be something like:

                my @entries = $doc->getBibliographyElements("^GEN99$");
                foreach my $entry (@entries)
                                author  => 'Genicorp',
                                pages   => 62

        Caution: This method allows the user to create any new property and
        to put any value in any property, without control. So we recommend
        you to have a look at the Open Office XML specification and/or the
        OOo bibliographic project ( in
        order to know the generally accepted properties.

bookmarkElement(element, name [, offset])

        Puts a bookmark in a text element.


                my $paragaph = $doc->selectElementByContent
                                                ("Eragon and Saphira");
                $doc->bookmarkElement($paragraph, 'Mark');
        puts a bookmark identified by "Mark" in a paragraph where a given
        text has been found (of course, the bookmark will remain even if
        the text of the paragraph is changed later).

        By default, the bookmark is put at the beginning of the text. But,
        thanks to the optional offset, it can be put at any position within
        the text of the bookmarked element.

cellCurrency(table, row, column [, currency])

cellCurrency(cell [, currency])

        Get/set the currency unit of a cell.
        If a currency is provided, the cell value type is automatically
        switched to 'currency'.

cellFormula(table, row, column [, formula])

cellFormula(cell [, formula])

        Accessor which returns the formula (or function) contained in the
        given table cell. Returns undef if no formula is found in the cell.

        The cell address is the same as for getCellValue().

        If a formula is given as the last argument, it is put into the cell,
        overwriting any existing formula. No check of the syntax is carried
        out on the inserted formula. It is up to the application to insert a
        formula which conforms to syntax. Example:

            $doc->cellFormula(1,3,2, "sum <C2:C5>");

        Note 1: inserting or replacing a formula does not directly modify
        the value or text of the cell. Proper interpretation of a formula
        does not happen until the fields are updated when the document is
        reloaded into

        Note 2: syntax and functionality of cell formulae differ greatly
        between the Writer and Calc applications.

cellSpan(table, row, column [, span])

cellSpan(cell [, span])

        In a spreadsheet document, get/set the span of a table cell,
        knowing that this span can be one or more columns. The cell addressing
        is the same as with getCell().

                $doc->cellSpan($table, "B4", 3);

        creates a 3-cell span from B4 in a spreadsheet.

        This method works only for horizontal expansion.

        The text of the covered cells (if any) is concatenated to the original
        content of the expanded cell (as in OOo Writer or Calc).

        Caution: when related to table cells, "span" has not the same
        meaning as when related to flat text (see getSpan() and setSpan()).

cellStyle(table, row, column [, stylename])

cellStyle(cell [, stylename])

        Get or set the style of a table cell.

cellValue(table, row, column [, value [, text]])

cellValue(cell [, value [, text]])

        Without the "value" argument: see getCellValue().

        With "value" (and, optionnally, "text"): see updateCell().

cellValueType(table, row, column [, type])

cellValueType(cell [, type])

        Get/set the data type of a table cell.

        Note: If an application must convert a 'string' cell to a numeric
        one and fill it with a numeric value, cellValueType() must be called
        *before* cellValue(). Ex:

                my $cell = $doc->getCell('Sheet1', 4, 8);
                $doc->cellValueType($cell, 'float');
                $doc->cellValue($cell, 12.34);

columnStyle(column_element [, style])

columnStyle(table, column [, style])

        Returns the style name of the given column or replaces it with a new
        one. A column can be indicated either directly by reference or by
        the pair [table, column number]. The table itself can be indicated
        either by a table element, its number or its logical name. If the
        'style' argument is given, it replaces the old column style.

        Giving a column a style is actually the only way to control the
        width of a column in a table.


            $doc->columnStyle('Table1', 2, 'NewStyle');

        Caution: columns are numbered beginning at 0.


        Get or set the default terminator character for text export.


        After this instruction, a line-break will be appended at the end of
        every paragraph or header exported by getText(), selectTextContent()
        or other text extracting methods.

        To reverse this behaviour, the user can call this method with an
        empty string.

        Without argument, returns the currently selecter terminator, if any.


        Deletes the given bookmark (if defined).


        Returns the list of the bibliographic entry elements contained in the

        If an argument is provided, the returned list is restricted to the
        bibliographic entries matching it (this argument can be a regex).


                my @biblio = $doc->getBibliographyElements("^W3C");
        returns the bibliographic entries where the identifier begins with


        Returns the bookmark element (if defined) corresponding to the given
        bookmark name.

        If the bookmark covers a range of text (i.e. if it's not a position),
        the returned element is the "bookmark start" one.

getCell(table, row, column)

getCell(table, coord)

getCell(row, column)

        Returns the element which represents the given cell. Possible
        arguments are respectively: the table number or its reference in the
        document, row number and column number. Each table cell contained in
        the body of an document can be referenced in this
        manner, as if it belonged to a single 3D table irrespective of the
        rest of the document [22] .

        The first argument can be either the sequential number of the table
        (starting at 0), the logical name of the table, or a 'table' object
        (which can be retrieved in advance using getTable). If it's a number
        or a name, getTable() is automatically called by getCell() in order
        to convert it in a 'table' object. However, if the first argument is
        a row object (previously obtained via getRow()), the second one is
        processed as the column number. Before using several cells in the
        same row, it's a good idea to get the row object and then to use it
        in every cell selection, in order to minimize the coordinates

        Alternatively, the user can provide the cell coordinates in a single
        alphanumeric argument, beginning with one or two letters and ending
        with one or more decimal digits, according to the same logic as in a
        spreadsheet. So, for example

                $doc->getCell($table, 'B12');
        is equivalent to

                $doc->getCell($table, 11, 1);

        (Remember that, with the numeric coordinates, the row number is the
        first argument, while with the alphanumeric, spreadsheet-like ones,
        the column letter(s) come first.)
        Numbers can also be negative, where position -1 is the last. For

            $cell = $doc->getCell(-1, -1, -1);

        returns the very bottom right cell of the very last table in the
        document $doc.

        Returns a null value if the given cell does not exist or if it's
        covered by the span of another cell.

        Any cellXXX() method in this module uses the same cell addressing
        logic as getCell().

        Note about spreadsheets:
        Addressing cells in spreadsheets is considerably more complex
        than in text document tables. However, the same addressing scheme
        in allowed in the "Calc" documents than in the "Writer" ones,
        provided the targeted cells belong to a "managed area" (beginnning
        to the upper-left cell, and ending at a parametrizable position).
        If the 'expand_tables' property of the document is set to 'on', this
        managed area is automatically initialized when the sheet is
        targeted for the first time. So, the first access to a given sheet
        (whatever the row or the cell) can be significantly more costly
        than any subsequent access. See normalizeSheet() for more
        explanations. Remember that the table addressing is zero-based and
        the row comes before the column in OpenOffice::OODoc, so, for

                $cell1 = $doc->getCell($table, 0, 0);
                $cell2 = $doc->getCell($table, 31, 25);

        returns respectively the A1 and Z32 cells.

        Note: in a spreadsheet, (0,0) are the coordinates of the "A1" cell,
        and, for example, (16, 25) are the coordinates of the "Z17" cell.

getCellValue(table, row, column)


        Returns the value of a table cell (and not the cell element as with

        The first form indicates a cell by its 3D coordinates, as with

        The second form (quicker) takes a cell element as its only argument
        (e.g. as returned by a previous getCell call).

        This method behaves in two different ways depending on the cell

            - returns the cell's text if the cell is set to literal (after
            any UTF8 decoding. See OODoc::XPath).

        This difference in handling is designed to allow programs to use
        returned numeric values directly in calculations.

        Note: To get information about a cell other than its value (numeric,
        etc.), the best way is first to get its element reference with
        getCell and then use it with getAttribute.

getChapter(header_no [, options])

        This method returns the list of the elements depending on a given
        header element. The argument and the options are the same as with


                my @list = $doc->getChapter(2, level => 3);
                foreach my $element (@list)
                        my $text = $doc->getText($element);
                        print "$text\n";

        The code above selects and prints all the text elements below the
        third level 3 header of the document (not including the content of
        the header itself.

getColumn(table, column)

        Returns the element reference of the given column in the given
        table. The first argument is either the table's sequential number in
        the document, logical name or element reference. The second argument
        is the column's number in the table. Synonym: getTableColumn.


        Returns the list of all the footnote citations (i.e. references to
        footnotes included in the text) contained in the document.


        Returns the list of all the footnote elements contained in the

getHeader(n [, options])

        Returns the nth+1 header element.

        If n is negative, headers are counted backwards from the last.
        getHeader(-1) returns the last header element of the document.

        The only one possible option is "level". It allows the application
        to select the nth+1 header element for a given header level.


                my $header = $doc->getHeader(2, level => 3);
        selects the third level 3 header in the whole document.

        See also getChapter().

        Caution: without the "level" option, this method counts sequentially
        through all headers along a single plane, irrespective of their
        level[24]. E.g. if you have a level 1 header then two level 2 headers
        then a level 1 header, the call getHeader(3) returns the last level 1


        Returns a list of header elements (i.e. elements called 'text:h' in
        the document body).

        This list actually contains the information necessary to create a
        contents list of an document, from within a document
        management application.


        Returns the text of the nth+1 header element. Elements are counted
        in the same way as for getHeader.


        Returns a list of document header texts.

        An optional filter argument can be passed (literal or regular
        expression). In this case, only headers whose content matches the
        filter are returned.

        In a list context, the result is returned in the form of a list of
        character strings. In a scalar context, the result is a single
        string in which the headers are separated by a line-feed character

        Note: This list is "flat". It contains no information about the
        headers' hierarchy. To get a hierarchical contents list, you must
        start with the list of headers obtained using getHeaderList and
        check each element's level attribute ('text:level').


        Returns a list of elements which represent items of an ordered or
        unordered list. The argument is a "list" element (obtained
        previously e.g. using getItemList, getOrderedList or
        getUnorderedList). Each element in this list can be used with item
        handling methods.


        Returns the element which represents the nth+1 ordered list in a
        document if found.

        WARNING: In the 1 documents, only "ordered lists" and
        "unordered lists" can be present. In the Open Document format, there
        are generic list objects only, and each one is made "ordered" or
        "unordered" by its style. So, this method will never return anything
        from an OOo 1 document.


        Returns the element which represents the nth+1 ordered list in a
        document if found.

        WARNING: Ordered lists are possible in the 1 format
        only. Don't use it against Open Documents.


        Returns the nth+1 paragraph in the document body, or undef if the
        given number is greater than or equal to the total number of
        paragraphs in the document.

        You can also pass a negative argument, in which case paragraphs are
        counted backwards from the end (-1 being the last paragraph).

        By paragraphs we mean 'text:p' elements, which excludes headers but
        includes non-empty table cells, contents of list items and

        Returned value is an element and not the text of the paragraph. All
        read/write operations involving attributes and content can use this


        Returns a list of paragraph elements (i.e. 'text:p' elements in the
        document body).


        Returns the text of the nth+1 paragraph, counted using the same
        rules as for getParagraph.


        Returns a list of texts contained in the paragraphs of a document
        ('text:p' elements).

        A filter can be passed as an optional argument (literal or regular
        expression). In this case, only paragraph texts whose content match
        the filter are returned.

        In a list context, the result is returned in the form of a list of
        character strings. In a scalar context, the result is a single
        string in which the paragraphs are separated by a line-feed
        character ("\n").

getRow(table, row)

        Returns the element reference which corresponds to a row in a table.
        The first argument is either the table's sequential number in the
        document, logical name or element reference. The second argument is
        the row number in the table. Synonym: getTableRow.

getRowCells(table, row) =head3 getRowCells(row)

        Returns the list of the cell elements corresponding to a given table row.
        The row can be provided either by table ID and row number or by direct
        row object.


        Returns a list of elements which correspond to texts which "stand
        out" from the document i.e. which have been given a style which
        makes them stand out from the rest of the paragraph containing them.

        For example, a word in italics or in font size 12 in a paragraph of
        mostly standard characters in font size 10 is a 'span' element and
        would therefore appear in a list returned by getSpanList.


        Gets a list of texts which "stand out" in the same way as
        getSpanList and returns it under the same conditions as
        getParagraphTextList or getHeaderTextList, with optional filter.

getStyle(path, position)


        Obsolete. See textStyle.

getTable(n [, length, width])

getTable(name [, length, width])

        Returns the nth+1 table in a document, if found, or undef if not
        found [25] .

        The second form allows you to select a table by its logical name (as
        it would appear to the end user when editing the table's
        properties). This name is obviously easier to use than a number.
        Moreover, this type of selection means the application will still
        work even if a table changes position within a document.

        The returned object is a "handle" that can be used for subsequent
        accesses to its components (rows, cells).

        getTable() can be used to retrieve a sheet in a Calc document as
        well as a table in a Writer document. However, before using any of
        the row/column/cell manipulation available methods, a special
        preprocessing should be done if the target table is a spreadsheet.
        See normalizeSheet() for more information.

        A getTable() call with the optional length, width arguments produces
        the same effect as an explicit call of normalizeSheet() with the same

        In the text documents, the tables are used without any preprocessing
        and the paragraph above doesn't apply.

        The returned value is a table element and not a table's content.

getTableColumn(table, column)

        See getColumn.


        Returns a list of table elements in a document.

getTableRow(table, row)

        See getRow.


        Returns the list of the rows contained in the given table.

        When the user needs to process every row in large tables, this method
        allows some performance improvements, because it's less costly than
        a lot of successive getRow() calls.


        Returns the size of a table as a pair of values which represent the
        number of rows and columns. The table can be specified either by
        number, logical name or reference.


            my ($rows, $columns) = $doc->getTableSize("Table1");


        Returns the content of a table, if found, whose number or reference
        is given as an argument. If not found, returns undef.

        The content of each cell is extracted according to the rules of

        In a list context, the returned value is a 2D table with each
        element containing the corresponding cell in the document.

        In a scalar context, the content is returned as a single string in
        CSV format. In this case, the rows are separated by a delimiter set
        by the instance variable 'line_separator' and the fields by the
        variable 'field_separator' in the OODoc::Text object. (These
        delimiters are by default "\n" and ";" respectively.)

getText(path, position)


        Exports the text contained in the given element according to the
        means appropriate to that type of element.

        If the 'use_delimiters' flag is set to 'on' (default), the content
        of each element (others than ordinary paragraphs, table cell,
        headers) is preceded and/or followed by a character string depending
        on the type of the element. This also depends on the settings given
        to the delimiter values 'begin' and 'end' by the 'delimiters' hash.
        In a default configuration where the application has not provided
        any specific delimiters, the following delimiters are used:

            - '<<' before and '>>' after sections of text highlighted within
            an element (e.g. words in bold or underlined within a paragraph
            of 'standard' font characters).

        footnote citations (in text body) are placed between square

        '{NOTE:' and '}' for the content of footnotes [26] .

        An application can change these delimiters, add more for other types
        of elements (e.g. paragraphs, headers, tables cells, etc.), or
        deactivate them using outputDelimitersOff. This depends on where the
        text is exported to e.g. display in editable "flat" format,
        conversion to XML or a markup language other than
        XML, generating code from text, etc..

        A default export (ex: "\n") terminator can be set for any element that
        is not listed in the 'delimiters' hash (see defaultOutputTerminator()

        If the element is an ordered or unordered list, the text produced is
        a concatenation of all the lines in the list, each separated by a
        line-feed ("\n") in addition to any delimiters.

        If the element is a table cell, getText behaves like getCellValue.

        If the element is a table, getText behaves like getTableText.


        Returns the text of a document, as "flat" editable text.

        In a list context, the content is returned as a table with one text
        element (header or paragraph) per element.

        In a scalar context, the content is returned as a single character
        string with each text unit (header or paragraph) separated by a
        line-feed ("\n").

        The returned text contains no style or level information, so there
        is nothing to distinguish a header from a paragraph.

        Same as selectTextContent('.*').


        Returns the list of all the text elements, including headers,
        paragraphs and item lists.


        Same as getParagraph but only considers top level paragraphs. The
        contents of lists, tables and footnotes are excluded.


        Returns the element which represents the nth+1 unordered list in a
        document, if found.

        WARNING: Ordered lists are possible in the 1 format
        only. Don't use it against Open Documents.


        Returns the element (if defined) representing a user-defined field,
        and corresponding to the given name. See also userFieldValue().


        Returns the UTF8 conversion of the given text, supposed to be in
        the local character set of the document (see the 'local_encoding'

insertHeader(path, position, options)

insertHeader(element, options)

        Same as appendHeader, but inserts the given header at the given

        Position is that of an existing element which can be another header
        or a paragraph. Can be given by [path, position] or by element

        Possible options are the same as for appendHeader, with the
        additional option 'position' which determines if the header is
        inserted before or after the element at the given position. Possible
        values for this option are 'before' and 'after'. By default, the
        element is inserted before the given element.

insertItemList(path, position [, options])

insertItemList(element [, options])

        Same as appendItemList, but a new list is inserted at the given
        position. The point of insertion can be given either by the pair
        [path, position] or by element reference. Options are the same as
        for insertParagraph.

insertParagraph(path, position [, options])

insertParagraph(element [, options])

        Same as appendParagraph, but a new paragraph is inserted at the
        given position.

        Position is that of an existing element which can be another
        paragraph or a header. Can be given by [path, position] or by
        element reference.

        Options are the same as for appendParagraph, with the additional
        option 'position' which determines whether the paragraph is inserted
        before or after the element at the given position. Possible values
        for this options are 'before' and 'after'. By default, the element
        is inserted before the given element.

insertRow(table, row [, options])

insertRow(row_element [, options])

        Inserts a new row into a table. In its first form, pass the table
        (reference, logical name or number) and the position number in the
        table. In its second form, pass the element reference of the
        existing row which is directly before or after the position where
        you want to make the insertion.

        By default, the new row is inserted at the position of the
        referenced row, which displaces it and the rest of the table down by
        one row position. However, you can insert it after by using the
        'position => after' option. By default, the new row is an exact copy
        of the referenced row, but you can assign particular attributes to
        it in the same manner as the insertElement method of OODoc::XPath.

insertTable(path, position, name, rows, columns [, options])

insertTable(element, name, rows, columns [, options])

        Creates a new table and inserts it immediately before or after
        another element (paragraph, header, table). The referenced element
        can be indicated as in insertParagraph. The other arguments and
        options are the same as for appendTable with the additional option
        'position' as in insertParagraph.

insertTableRow(table, row [, options])

insertTableRow(row_element [, options])

        See insertRow.

insertText(path, position, element_name, options)

insertText(element, name, options)

        As appendText, but a new text element is inserted at the given

        The position is that of an existing element (of any type). It can be
        given by [path, position] or by element reference.

        Options are the same as for appendText, with the additional option
        'position' which determines whether the element is inserted before
        or after the element at the given position. Possible values for this
        option are 'before' and 'after'. By default, the element is inserted
        before the given element.


        This module implements a set of accessors which can be used like
        native text element methods as opposed to OODoc::Text methods, and
        which allow you to determine what type a given element is [27] . For
        an element $e of document $d, the syntax is $e->isXxx instead of $d-


            print "This is a list" if $element->isItemList;

        Here is the list of element type indicators:

            isFootnoteBody              footnote

            isFootnoteCitation  footnote citation

            isHeader                    header

            isItemList          list (ordered or unordered)

            isListItem          list item

            isOrderedList               ordered list

            isParagraph         paragraph

            isSequenceDeclarations      set of sequence declarations

            isTable                     table

            isTableCell         table cell

            isUnorderedList             unordered list

        For a neater and more direct access to element types, see the
        getName method of XML::Twig [28] .

normalizeSheet(sheet [, rows [, columns]])

        To be used with spreadsheets. This method preprocesses a given
        sheet so its components (rows, cells) become available for all the
        table-oriented methods described in this chapter.

        This method is not needed for tables included in
        Writer (sxw) documents, because these tables are "normalized" (i.e.
        each component is mapped to an exclusive XML element). But in Calc
        (sxc) documents, the XML mapping of rows, columns and cells is
        "denormalized" in order to save memory: several table components
        can be mapped to a single XML element, so the XML address of each one
        can't be simply calculated from its logical coordinates (sheet, row,
        column). In order to allow the spreadsheets components to be addressed
        with the same methods as the Writer table components, normalizeSheet()
        reorganizes the XML mapping of the given sheet.

        Because this method is very time and memory consuming, it should never
        be used to reorganize the largest possible area of a sheet (meaning
        thousands of rows and hundreds of columns that will probably never be
        used). So it's action is limited to a given area, controlled by the
        rows, columns arguments. When these arguments are not provided, the
        method uses the 'max_rows' and 'max_cols' properties instead (see the
        Properties section for other explanations).

        The first argument can be either the logical name of the sheet (as
        it's shown in the bottom tab by OOo Calc), the sheet number, or a
        table object reference, previously returned by getTable(). The return
        value is the table object (or undef in case of failure).

                $doc = ooDocument(file => 'report.sxc');
                my $sheet = $doc->normalizeSheet('Sheet1', 7, 9);
                my $result = $doc->cellValue($sheet, 5, 6);
        In the sequence above, a top left area of 7 rows by 8 columns is
        pre-processed, so the cells from A1 to H6 of this sheet can be
        reached according to the same addressing scheme as in Writer tables.
        The last instruction gets the content of G6.

        Because a "normalized" sheet has the same XML structure as a Writer
        table, it's generally possible to directly copy it from a spreadsheet
        document to a text document. Example:

                $doc1 = ooDocument(file => "spreadsheet.sxc");
                $doc2 = ooDocument(file => "text.sxw");
                $sheet = $doc1->normalizeSheet("Sheet1", 6, 8);
        In this last example, a new table, that is a copy of the A1:H7 area
        of the "Sheet1" sheet of a Calc document, is attached at the end of
        a Writer document.

        The transformed sheets, of course, are readable by OOo Calc.
        They simply take some more disk space when the processed spreadsheet
        is saved. If the document is later read then written by OOo Calc,
        the storage is optimized, so the effects of normalizeSheet()

        normalizeSheet() can be used safely against Writer document tables,
        with two possible results. If the table size is greater than the given
        size, the method is neutral. Otherwise, the length and/or the width is
        increased according to the given arguments.

        An explicit call to this method can be replaced by getTable() with the
        additional length and width parameters. 



        Turns delimiters on or off. Used to mark up text exported by certain
        methods like getText or selectTextContent.

        The delimiters actually used depends on the table loaded into the
        OODoc::Text instance via the 'delimiters' property.


        Returns the conversion in local character set of the given text,
        supposed to be in UTF8. The local character set of the document
        is used (see the 'local_encoding' property).


        See deleteBookmark().



        Removes the header at the given position (first form).



        removes the 5th header (whatever its level) counted from the
        beginning of the document.

        The header to be removed can be indicated by element reference
        (second form). In this case, the type of element is not checked and
        this method becomes the equivalent of removeElement.



        Removes the paragraph at the given position (first form).

        The paragraph to be removed can be indicated by element reference
        (second form). In this case, the type of element is not checked and
        this method becomes the equivalent of removeElement.


        Removes the multi-column span of a table cell. The width of the cell
        is reduced to the width of its column. The uncovered cells take the
        same style and data type as the reduced cell.

removeSpan(path, position)


        "Flattens" a text element, removing all presentation distinctions
        which may mark out some sections of its content.

        For example, the paragraph:

            I consider Writer to be my word processor of


            I consider Writer to be my word processor of

        when this method is applied.

        See also setSpan.

rowStyle(row_element [, style])

rowStyle(table, row [, style])

        Reads or modifies a table row's style, in the same way as
        columnStyle does for columns.


        Returns the element containing the given bookmark.

selectElementByContent(filter, [...])

        Returns the first text element whose content matches the 'filter'
        (which can be an exact string or a regular expression), or undef
        if no matching content is found.

        With more than one argument, this method can be used for replacement
        operations, or user-defined function triggering, in the same
        conditions as selectElementsByContent.


selectElementsByContent(filter, replacement)

selectElementsByContent(filter, action [, other_arguments])

        This method returns a list of text elements such as paragraphs,
        headers or ordered/unordered lists whose content matches the search
        criteria contained in 'filter' (which can be an exact string or a
        regular expression).

        The first form simply returns the given list without modifying the

        The second form returns the same list, but replaces all strings
        which match the search criteria with the 'replacement' string as it

        The third form, where the 'action' argument is a program function
        reference, launches the given function each time the filter string
        is matched. If defined, the value returned by the function is used
        as the replacement value [29] . If the function returns a null value
        (undef) then no replacement is made. If it returns an empty string,
        the retrieved text is deleted. The called function receives the rest
        of the arguments, in this order:

        1) all remaining arguments after 'action' ('other_arguments'), if any.
        2) the element containing the retrieved text.
        3) the string actually selected. If the filter is an exact string,
        it is equal to the filter. If the filter is a regular expression,
        it matches the "real" text retrieved.

        The returned text (if any) must be encoded in UTF8.

        The returned list is the same one returned by the first two forms.


            sub action
                my ($d, $element, $value) = @_;
                if ($value < 100)
                        return undef;
                        return $value * 2;
                        @list =
             $doc->selectElementsByContent("[0-9]+", \&action, $doc);

        In the above code, the subroutine "action" is called each time an
        integer (one or more digits) is found. The subroutine receives the
        document reference itself as its first argument (an OODoc::Text
        object given by the application). Next, it automatically receives
        the reference of the element in which the search string was found
        (i.e. an integer) and, finally, it receives the exact number found
        as its second-last and last arguments respectively. If this number
        is less than 100, the element is removed. This is why the subroutine
        needed the $doc object, used to invoke the removeElement method. If
        more than 100, the number is multiplied by two and the result
        replaces the original value in the element. The list returned by
        selectElementsByContent contains all elements which contain the
        search string, including any which might have been removed by the
        called function while it was running.

        It is the "main" elements containing strings which matched the
        filter which are returned and not any of their sub-elements. For
        example, if the returned string is found in one of the items in an
        unordered list, the list element is selected and not the item.
        Similarly, the table is selected when one of its cells matches the
        filter, and the paragraph which is selected when the search string
        is found in an attached footnote.

        However, a character string cannot be considered to match the filter
        unless it is entirely within the same sub-element and all its
        characters have the same style. For example, if you were searching
        for the string "OpenOffice" using selectElementsByContent, the
        following paragraph would not be selected:

        This document was created using 1.0.1

        In this example, the word "OpenOffice" is there but it is divided by
        a style change which the logic of selectElementsByContent treats as
        belonging to two separate sub-elements. [30]  (This is not the same
        logic applied by the search function of an office application.)

        Note: This method can be used with a "non-filtering" regular
        expression (".*") for unconditional movement through all text


        Returns the first paragraph (if any) using the given style.


        Returns the list of the paragraphs using the given style.


selectTextContent(filter, replacement)

selectTextContent(filter, action [, other_arguments])

        Returns a list of header texts and/or paragraphs (in the document's
        own order) which match the given search criteria.

        The filter can be an exact string or a regular expression. A filter
        set to ".*" (no selection) will result in an export of the entire

        In all three forms, this method behaves like
        selectElementsByContent, except that it returns text instead of a
        list of elements.

        Depending on the context (list or scalar), the result is returned in
        the form of a list of rows or in the form of a single character
        string where the elements are separated by a line-feed ("\n").

        Note: called with a "non-filtering" regular expression, this method
        will result in a "flat" export of the document:

            print $doc->selectTextContent('.*');

setSpan(path, position, [context,] expression, style)

setSpan(element, [context,] expression, style)

        Applies a "span" to part of the content of a text element.

        In XML language, a "span" is a substring whose
        presentation style differs from the style of the text element to
        which it belongs. For example, a given "span" could be in italics
        while the rest of the paragraph is in normal characters.

        Caution: the same word has a different meaning when it's used
        about table cells (see cellSpan()).

        A "span" is therefore a way to use several styles within the same
        element, bearing in mind that the paragraph's global style can be
        modified by setStyle.

        The properties of a span can be related to any kind of character
        string presentation, such as font, font size, font weight, font
        style, and colors (background and foreground).

        The desired text element is normally indicated by [path, position]
        or reference. The optional argument 'context' which consists of an
        element reference, allows you (when using [path, position]) to limit
        a search to child elements of a particular element (e.g. headers,
        footers, unordered lists, etc.).

        'expression' is the character string to be highlighted, and 'style'
        is obviously the style describing the presentation characteristics
        to give to it. See OODoc::Styles for how to construct styles from

        As a highlighted string can be quite long or not all known in
        advance, you can represent it with a regular expression. Taking the
        previous example again, you could use:

            my @list = $doc->selectElementsByContent("OpenOffice");
            foreach my $para (@list)
                $doc->setSpan($para, "Open.*Writer", "Style1");
                $doc->setSpan($para, "word.*ssor", "Style2");

        This sequence selects all text elements (or, more simply,
        paragraphs) which contain "OpenOffice". In each of them, the style
        "Style1" is given to the string beginning with "Open" and ending
        with "Writer" (if found) and the style "Style2" is given to the
        string beginning with "word" and ending with "ssor". Defined
        elsewhere, Style1 and Style2 would correspond to the required
        presentation characteristics.

        Caution: the current version of this method can neither recognise
        nor handle a string located partly in a "span" and partly outside
        it. It can, however, create a "span" inside another.

        See also removeSpan.

setStyle(path, position, style_name)

setStyle(element, style_name)

        Obsolete. See textStyle.

setText(element, text ,[text, ...])

        Alters the setText method of OODoc::XPath, so that it can handle
        complex text elements.

        If the element is a paragraph, a header of a list item (ordered or
        unordered), its content is replaced by the 'text' argument. Caution:
        setText deletes and replaces the previous content of the paragraph.

        If the element is a table cell, this method is the same as

        If the element is a list (ordered or unordered), the content of each
        'text' argument (however many) forces the creation of a new item
        which is appended to the list (existing items remain unchanged).

            $doc->setText($element, "Peter", "Paul", "John")

        adds three items to the list if $element is a list. If $element is,
        for example, a paragraph, then the second argument ("Peter") becomes
        the content of the paragraph and the other arguments are ignored.

        For all other types of element, setText behaves normally as defined
        in OODoc::XPath.

tableStyle(table [, style])

        Returns the current style of a given table, or replaces it with a
        new style given as the second argument. The table can be indicated
        by number, logical name or reference.

textStyle(path, position [, style])

textStyle(element [, style])

        Reads a text element's style or, if a 'style' argument is given,
        changes it.

        The element can be indicated by the pair [path, position] or by

        Note: the returned value is a literal style identifier or the value
        of the element's 'text:style-name' attribute.

        Note: this method allows you to attribute a non-existent style to a
        paragraph or header. Such a style can be created later (e.g. using
        createStyle) or not at all. The actual existence of the style is
        only relevant to the needs of the application [31] .

updateCell(table, row, column, value [, text])

updateCell(element, value [, text])

        Modifies the content of a table cell.

        In its first form, indicates a cell by its 3D coordinates, as with
        getCell(). In its second form, indicates a cell by its element

        If the cell is set to literal, its content is limited to its text.
        In this case, the optional argument "text" is of no use (the text
        equals the value).

        If the cell is set to numeric (float, currency, date, etc.), you
        should generally pass a literal argument as well as the value.

        This method can be replaced by the accessor cellValue which allows
        reads and writes.

userFieldValue(user_field [, value])

        Reads the stored value of a given user field or changes it if a
        value is provided. The 1st argument can be either the name of the
        field (as it appears for the end-user) or a previously loaded
        user field element. See also getUserFieldElement().

        This method doesn't create any new user field. It can only read or
        update an existing one.

        If the given user field is numeric (ex: date, currency) the returned
        and/or provided value is the internally stored value, and not the
        displayed one. The user field is displayed according to a data style
        by For example, 'Tuesday, March 1, 2005' is a possible
        displayed value for 38412.


        No class variables are exported.

        Instance properties are the same as for OODoc::XPath, plus:

            'delimiters'        => delimiter table

        hash giving the relation between element types and the delimiters to
        use when exporting text (see getText).

            'use_delimiters'    => delimiter usage (see getText)

        indicates whether delimiters are to be used by getText or not when
        exporting text. Set to 'on' by default. Can be set to 'off' or
        another value to stop or limit use of delimiters.

            'header_style'      => default header style

        indicates the default header style to be used by element creation
        methods when no style is specified. Set to 'Heading 1' by default.

            'paragraph_style'   => default paragraph style

        indicates the default paragraph style to be used by element creation
        methods when no style is specified. Set to 'Standard' by default.

            'field_separator'   => field separator

        contains the character string to be used as the field separator when
        exporting tables. By default it is ";".

            'line_separator'    => line separator

        contains the string to be used to separate lines when exporting
        "flat" text. By default, it is a line-feed ("\n").

            'max_rows'          => max table length (default 32)
            'max_cols'          => max table width (default 26)

        these 2 properties control the size of the "managed area" in a
        spreadsheet; the default "managed area" is the A1:Z31 rectangle,
        corresponding to the (0,0)-(31,25) coordinates; see getTable() and
        getCell() and normalizeSheet() for more explanations.

            'expand_tables'     => table transformation usage

        indicates whether the XML representation of the spreadsheets are to
        be expanded in order to allow the same cell/row addressing scheme
        as with the tables belonging to text documents; by default, this
        property is not set. If this property is set to 'on', the first
        access to any sheet will automatically trigger this transformation,
        so the explicit normalizeSheet() method will not be needed.
        However, this automatic (but costly) transformation has a drawback:
        it uses the same 'max_rows' and 'max_cols' values for every targeted
        sheet, whatever the really needed managed area for each one.


See OpenOffice::OODoc::Notes(3) for the footnote citations ([n]) included in this page.


Copyright 2004-2005 by Genicorp, S.A. (

Initial developer: Jean-Marie Gouarne (

Initial English version of the reference manual by Graeme A. Hunter (


        - Genicorp General Public Licence v1.0
        - GNU Lesser General Public License v2.1