Handling spaces and whitespace in XML

XML Extensions normally strips trailing spaces from COBOL data items when exporting data and restores trailing spaces to COBOL data items when importing data. Leading spaces are also removed and added for justified data items.

This default behavior can be modified using the XML SET FLAGS statement, but the default behavior is generally best. The normal treatment of leading and trailing spaces does not apply to FILLER data items or edited data items.

Once the data is in XML, further consideration must be given to XML treatment of whitespace, which includes spaces, carriage returns, and line feeds. XML provides a built-in attribute named xml:space, which takes a value of "preserve" or "default". The value "preserve" specifies that whitespace in an element should be preserved. The value "default" specifies that leading and trailing whitespace may be removed and embedded whitespace may be normalized to a single space wherever it occurs. The value "default" is the default treatment of whitespace in XML and is generally not changed unless one is trying to produce poetry or other special output.

XML Extensions has various rules for handling of the individual whitespace characters during import or export as follows:

• Space characters are preserved by XML Extensions during import and export, except for leading and trailing spaces as described in the first paragraph of this topic.
• During import, all characters received from the parser for XML text nodes for import are preserved. Preserving all characters is the default behavior, unless the XML SET WHITESPACE-FLAGS statement is used to change the whitespace flags from their default value of 0. Since the introduction of XML SET WHITESPACE-FLAGS, this is a change in behavior. Before this new behavior, characters less than space, which include line feed, carriage return and tab characters, were unconditionally removed on import. The new statement XML SET WHITESPACE-FLAGS can set various whitespace handling behaviors, one of which is the old behavior if that behavior is desired.
• During export, line feed, carriage return and tab characters present in the exported COBOL data item are preserved in the exported text node. Export to a file may cause LF characters to be translated to a CR/LF sequence on some operating systems (for example, Windows).

The value, "preserve" or "collapse" of the XML reserved attribute XML-space is ignored by XML Extensions unless the parser acts on this attribute.

When using XSLT stylesheets, the xsl:strip-space and xsl:preserve-space elements indicate how whitespace should be handled while transforming a document. Preserving whitespace is the default, but tools that generate XSLT stylesheets might insert xsl:strip-space elements.

Be aware that when documents are transformed to HTML for display by a browser, many browsers strip whitespace as they are allowed to do. Displaying data in tables is generally necessary to align data in columns rather than using whitespace as is generally done in COBOL report output without XML support.