Improving the output capabilities of Stata with Open Document Format xml

1
Improving the outputcapabilities of Stata
withOpen Document Format xml

• Adam Jacobs
• Dianthus Medical Limited

2
Statas 3-fold capabilities

• Statistics
• Graphics
• Data management

3
Statistics
4
Graphics
5
Data management
6
But there is a 4th...
7
Text output

• A recent clinical study
• 92 pages of raw data listings
• 124 pages of descriptive data tabulations
• 3 pages of statistical analysis
• All from a study in 12 healthy volunteers

8
Statas text output
9
Problems with Statas text output

• No pagination
• No formatting (or limited formatting with smcl)
• Variable labels not always shown
• No Unicode support
• No tables of contents
• etc etc

10
Some examples...
11
So how did I do it?
12
Open Document Format

• An open standard, approved by ISO
• XML based
• For a variety of office-type documents
• Used by the popular open-source office suite
  OpenOffice.org
• Here, we are just interested in word-processing
  documents

13
.odt files

• A .odt file is the native file format of
  OpenOffice.org Writer
• A zip file
• Contains various files, the most important of
  which is content.xml
• content.xml is simply a plain-text file
• Stata is good at writing plain-text files!

14
The Stata code

• Creates the content.xml file by writing data with
  appropriate xml tags
• Added to other files, zipped to .odt file
• .odt file can be opened directly with Writer

15
Some examples...
16
Basics of XML

• ltcompany nameDianthus Medical Limitedgt
• ltemployee rolespeakergt
• ltfirstnamegtAdamlt/firstnamegt
• ltlastnamegtJacobslt/lastnamegt
• lt/employeegt
• ltemployee roledelegategt
• ltfirstnamegtFlavialt/firstnamegt
• ltlastnamegtWhitelt/lastnamegt
• lt/employeegt
• lt/companygt

17
XML code for start of table

• lttabletable tablestyle-name"Table42"gt
• lttabletable-column tablestyle-name"TabC
  ol13"/gt
• lttabletable-column tablestyle-name"TabC
  ol9"/gt
• lttabletable-column tablestyle-name"TabC
  ol8"/gt
• lttabletable-column tablestyle-name"TabC
  ol8"/gt

18
XML code for table cells

• lttabletable-cell tablestyle-name"cell1211"gt
• lttextp textstyle-name"Table_20_Contents"gt Mil
  eage (mpg)lt/textpgt
• lt/tabletable-cellgt
• lttabletable-cell tablestyle-name"cell1111"gt
• lttextp textstyle-name"Table_20_Contents"gtNlt/te
  xtpgt
• lt/tabletable-cellgt
• lttabletable-cell tablestyle-name"cell1111"gt
• lttextp textstyle-name "Table_20_ContentsNumer
  ic"gt
• 52lttexts textc"3"/gtlt/textpgt
• lt/tabletable-cellgt

19
Was this a lot of work?

• 123 kB of code
• 21 ado files
• 45 Mata functions
• And not finished yet!

20
Any questions?