SDA 3.5 Documentation for DDLTOX

NAME

ddltox - Convert a DDL file to SAS, SPSS, or Stata definitions or to XML (for DDI)

USAGE

ddltox [-option] -i input_file

DESCRIPTION

DDLTOX converts the data descriptions in a DDL file into data definitions for SAS, SPSS, or Stata. That file, together with a matching data file, can then be used to create a system file for one of those systems and to run data analysis procedures. The program can also generate variable definitions in XML, following the conventions of the Data Documentation Initiative (DDI- Version 2).

By default, DDLTOX creates SAS data definitions. To create definitions for SPSS, Stata, or XML, use the ‘-x’ option.

DDLTOX obtains the data description information from the DDL file given after the ‘-i’ flag on the command line. Since SAS, SPSS, and Stata do not define the attributes of variables in exactly the same way as DDL, there are occasionally some problems in converting from DDL to one of those systems. Some of the options of DDLTOX are designed to facilitate those conversions.

Warnings and error messages are placed in a file named ‘DDLTOX.MSG’. Users should view the contents of that file after running the program.

MEANING OF THE OPTIONS

Define Input and Output Files

-x system: Type of output to produce -- SAS, SPSS, Stata, or XML.
(Default is SAS; other systems must be specified either in capital letters or in lowercase.)
-i fname: Take input from the DDL file ‘fname’. (This specification is REQUIRED.)
-o fname: Write the new data definitions onto the file ‘fname’ (instead of to the standard output).

Options for Variable Names

-@ character(s)

Replace any ’@’ character in a variable name with the specified conversion character(s). Any ’@’ characters in category labels will also be converted, so that GOTO information will reflect the revised variable names.
(This is useful for dealing with names from CASES instruments, especially when converting to SAS or Stata definitions.)

-m number

Override the default maximum for the length of each variable name (currently 32 characters) and set it to the number given after ‘-m’. The new maximum must be between 8 and 32.

If a variable name exceeds the maximum length, the corresponding variable definitions are skipped, and a warning is placed in the ‘DDLTOX.MSG’ file.

-p prefix

Use the characters given after ‘-p’ as a prefix to each variable name. (This is useful if variable names begin with a number, and they are to be converted to SAS, SPSS, or Stata.)

Options for Category Labels

-s: Ignore short (bracketed) category labels -- SAS, SPSS, and Stata output only
(See the discussion on short/long category labels below.)
-n max_characters: Maximum number of characters to output as a short category label -- XML output only
(Default is 60)
(See discussion on short/long labels for XML output below. )

Other Options

-v fname: Limit the variables processed to those contained in the variable list file ‘fname’.
-h: Display short program help and available options. (The program will not do anything else.)

MISSING DATA CONVERSION

Each system has its own method of indicating which codes are to be considered invalid and therefore to be excluded from data analyses.

DDLTOX will attempt to convert as many missing-data specifications as it can for each variable. If there is something that cannot be converted, a warning is placed into the ‘DDLTOX.MSG’ file.

SAS missing-data specifications

All numeric missing-data specifications in the DDL file are converted to IF-statements in SAS. Each such statement sets the referenced variable to the value ‘.’, if the value in the data file matches the missing-data condition.

If character missing-data codes are used for numeric variables, a special missing-data statement is included which applies to all variables in the SAS definitions.

SPSS missing-data specifications

The first three discreet missing-data codes in the DDL for a variable are put into an SPSS ‘MISSING VALUES’ statement. Any additional missing-data or valid range specifications for numeric variables are converted to IF-statements.

Character missing-data codes for numeric variables are not converted. If any are encountered, a warning is given.

Stata missing-data specifications

All numeric missing-data specifications in the DDL file are converted to REPLACE-IF statements in Stata. Each such statement sets the referenced variable to the value ‘.’, if the value in the data file matches the missing-data condition.

Character missing-data codes for numeric variables are not converted. If any are encountered, a warning is given.

XML missing-data specifications

The DDI specification can handle all of the DDL missing-data conventions except for character missing-data codes for numeric variables. If any of those specifications are encountered, a warning is given.

SHORT/LONG CATEGORY LABELS

Short/long category labels for SAS, SPSS, or Stata (-s)

In a DDL file there can be a long label or text description for a response code, plus a short label given in square brackets. For example the category label specification could look like:

Consistently votes Democrat [Democrat].

where the long label or text is ‘Consistently votes Democrat’, and the short label is ‘Democrat’.

If a short label is given in the DDL file for this category
When generating specifications for SAS, SPSS, or Stata, DDLTOX will, by default, use the short category label instead of the longer category label.
If the ‘-s’ option is specified, however, DDLTOX will ignore any short labels in brackets and will always convert only the long category label for SAS, SPSS, or Stata. (Note that the ‘-s’ option has no effect when the output is an XML file.)
If a short label is NOT given in the DDL file for this category
If a short label is not given, DDLTOX will use up to one line of the long label (if there is one). If the long label has more than 60 characters, only the first 60 characters will be used.

Short/long category labels for XML (-n)

In the DDI specification, there are two XML specifications for the labels of categories -- the ’labl’ element, and the ’text’ element. In general, the ’labl’ element is intended to be used as a shorter label for statistical analysis programs, whereas the ’text’ element is intended to be used as a longer explanation of the meaning of a particular category.

In a DDL file the short version of a category label is put between square brackets after the longer category text. An example of such a label would be:

Definitely will vote in the next election [Definitely vote]

where the short label is ‘Definitely vote’.

If BOTH short and long labels are in the DDL file for this category
When converting such labels into XML definitions, the short label will be put out using the ’labl’ element, and the longer category text will be put out using the ’text’ element.
If there is no short (bracketed) label in the DDL file for this category
If the category label in the DDL file is less than or equal to a certain number of characters (default=60), it will be put out using the ’labl’ element, If the label is longer, it will be put out using the ’text’ element.
The ’-n’ option allows the user to define what is meant by ’shorter’ or ’longer’ when deciding whether to output the label with ’labl’ or ’text’. The number given after the option flag replaces the default dividing criterion of 60 characters.

EXAMPLES

ddltox -i myDDL -o mySAS -@_: Convert DDL to SAS definitions. Convert ’@’ in variable names to underscore symbols (’_’).
ddltox -x spss -i myddl -o mySPSS -s: Convert DDL to SPSS definitions. Do not use bracketed short category labels found in the DDL file.
ddltox -x xml -n 40 -i myddl -o myXML: Convert DDL to XML definitions following the DDI specification; category labels longer than 40 characters will be put out using the ’txt’ element instead of the ’labl’ element.

DDL	Data Documentation Initiative - Version 2
DDL	Data Description Language
xconvert	Convert SAS, SPSS, or Stata defintions into XML or DDL