Title stata.com

xmlsave — Export or import dataset in XML format

Syntax Menu Description Options for xmlsave

Options for xmluse Remarks and examples Also see

Syntax

Export dataset in memory to XML format

xmlsave filename



if

 

in

 

Export subset of dataset in memory to XML format

Main

doctype(dta) save XML file by using Stata’s .dta format

doctype(excel) save XML file by using Excel XML format

dtd include Stata DTD in XML file

legible format XML to be more legible

replace overwrite existing filename

xmluse options Description

doctype(dta) load XML file by using Stata’s .dta format

doctype(excel) load XML file by using Excel XML format

sheet("sheetname") Excel worksheet to load

clear replace data in memory

1

2 xmlsave — Export or import dataset in XML format

Menu

xmlsave

xmluse

Description

although sometimes you will want to explicitly specify which document type definition (DTD) to use

by typing

xmluse can read either an Excel-format XML or a Stata-format XML file into Stata. You type

Stata will read into memory the XML file filename.xml, containing the data after determining whether

the file is of document type dta or excel. As with the xmlsave command, the document type can

also be explicitly specified with the doctype() option.

Options for xmlsave

Main

doctype(dta), the default, specifies that an XML file will be exported using Stata’s .dta format

(see [P] file formats .dta). This is analogous to Stata’s binary dta format for datasets. All data

that can normally be represented in a normal dta file will be represented by this document type.

doctype(excel) specifies that an XML file will be exported using Microsoft’s SpreadsheetML

DTD. SpreadsheetML is the term given by Microsoft to the Excel XML format. Specifying this

legible adds indents and other optional formatting to the XML file, making it more legible for a person

to read. This extra formatting, however, is unnecessary and in larger datasets can significantly

increase the file size.

replace permits xmlsave to overwrite existing filename.xml.

Options for xmluse

is optional, use of doctype() is encouraged. If this option is omitted with xmluse, the document

type of filename.xml will be determined automatically. When this occurs, a note will display the

document type used to translate filename.xml. This automatic determination of document type is

not guaranteed, and the use of this option is encouraged to prevent ambiguity between various

SpreadsheetML is the term given by Microsoft to the Excel XML format.

sheet("sheetname") imports the worksheet named sheetname. Excel files can contain multiple

worksheets within one document, so using the sheet() option specifies which of these to load.

The default is to import the first worksheet to occur within filename.xml.

cells(upper-left:lower-right) specifies a cell range within an Excel worksheet to load. The default

range is the entire range of the worksheet, even if portions are empty. Often the use of cells()

is necessary because data are offset within a spreadsheet, or only some of the data need to be

loaded. Cell-range notation follows the letter-for-column and number-for-row convention that is

popular within all spreadsheet applications. The following are valid examples:

information is dictated by SpreadsheetML, there are no constraints to keep types consistent within

columns. When such inconsistent use of data types occurs in SpreadsheetML, the only way to

resolve inconsistencies is to import data as string data.

4 xmlsave — Export or import dataset in XML format

firstrow specifies that the first row of data in an Excel worksheet consist of variable names. The

default behavior is to generate generic names. If any name is not a valid Stata variable name, a

missing forces any inconsistent data types within SpreadsheetML columns to be imported as missing

data. This can be necessary for various reasons but often will occur when a formula for a particular

cell results in an error, thus inserting a cell of type ERROR into a column that was predominantly

of a NUMERIC type.

nocompress specifies that data not be compressed after loading from an Excel SpreadsheetML file.

Because data type information in SpreadsheetML can be ambiguous, Stata initially imports with

broad data types and, after all data are loaded, performs a compress (see [D] compress) to reduce

data types to a more appropriate size. The following table shows the data type conversion used

before compression and the data types that would result from using nocompress:

Remarks and examples stata.com

XML stands for Extensible Markup Language and is a highly adaptable text format derived from

SGML. The World Wide Web Consortium is responsible for maintaining the XML language standards.

See http://www.w3.org/XML/ for information regarding the XML language, as well as a thorough

definition of its syntax.

The document type dta, used by both xmlsave and xmluse, represents Stata’s own DTD for

representing Stata .dta files in XML. Stata reserves the right to modify the specification for this DTD

at any time, although this is unlikely to be a frequent event.

The document type excel, used by both xmlsave and xmluse, corresponds to the DTD developed

by Microsoft for use in modern versions of Microsoft Excel spreadsheets. This product may incorporate

When you import data from Excel to Stata, a common hurdle is handling Excel’s use of inconsistent

of Excel data. Stata, however, requires that all data in a variable be of one consistent type. This can

cause problems when a column of data from Excel is imported into Stata and the data types vary

across rows.

By default, xmluse attempts to import Excel data by using the data type information stored in

the XML file. If an error due to data type inconsistencies is encountered, you can use the options

However, identifying the problem and determining which option to apply can sometimes be difficult.

Often you may not care in what format the data are imported into Stata, as long as you can import

them. The quick solution for these situations is to use the allstring option to guarantee that all the

data are imported as strings, assuming that the XML file itself was valid. Often converting the data

back into numeric form after they are imported into Stata is easier, given Stata’s vast data management

commands.

To export the current Stata dataset to a file, auto.xml, type

To overwrite an existing XML dataset with a new file containing the variables make, mpg, and

weight, type