read.dta13: Read Stata Binary Files

Description

read.dta13 reads a Stata dta-file and imports the data into a data.frame.

Usage

read.dta13(file, convert.factors = TRUE, generate.factors = FALSE, encoding = "UTF-8", fromEncoding = NULL, convert.underscore = FALSE, missing.type = FALSE, convert.dates = TRUE, replace.strl = TRUE, add.rownames = FALSE, nonint.factors = FALSE)

Arguments

file

character. Path to the dta file you want to import.

convert.factors

logical. If TRUE, factors from Stata value labels are created.

generate.factors

logical. If TRUE and convert.factors is TRUE, missing factor labels are created from integers.

encoding

character. Strings can be converted from Windows-1252 or UTF-8 to system encoding. Options are "latin1" or "UTF-8" to specify target encoding explicitly. Stata 14 files are UTF-8 encoded and may contain strings which can't be displayed in the current locale. Set encoding=NULL to stop reencoding.

fromEncoding

character. We expect strings to be encoded as "CP1252" for Stata Versions 13 and older. For dta files saved with Stata 14 or newer "UTF-8" is used. In some situation the used encoding can differ for Stata 14 files and must be manually set.

convert.underscore

logical. If TRUE, "_" in variable names will be changed to "."

missing.type

logical. Stata knows 27 different missing types: ., .a, .b, ..., .z. If TRUE, attribute missing will be created.

convert.dates

logical. If TRUE, Stata dates are converted.

replace.strl

logical. If TRUE, replace the reference to a strL string in the data.frame with the actual value. The strl attribute will be removed from the data.frame (see details).

add.rownames

logical. If TRUE, the first column will be used as rownames. Variable will be dropped afterwards.

nonint.factors

logical. If TRUE, factors labels will be assigned to variables of type float and double.

Value

The function returns a data.frame with attributes. The attributes include

Details

If the filename is a url, the file will be downloaded as a temporary file and read afterwards.

Stata files are encoded in ansinew. Depending on your system's default encoding certain characters may appear wrong. Using a correct encoding may fix these.

Variable names stored in the dta-file will be used in the resulting data.frame. Stata types char, byte, and int will become integer; float and double will become numerics. R only knows a single missing type, while Stata knows 27, so all Stata missings will become NA in R. If you need to keep track of Statas original missing types, you may use missing.type=TRUE.

Stata dates are converted to R's Date class the same way foreign handles dates.

Stata 13 introduced a new character type called strL. strLs are able to store strings up to 2 billion characters. While R is able to store strings of this size in a character vector, the printed representation of such vectors looks rather cluttered, so it's possible to save only a reference in the data.frame with option replace.strl=FALSE.

In R, you may use rownames to store characters (see for instance data(swiss)). In Stata, this is not possible and rownames have to be stored as a variable. If you want to use rownames, set add.rownames to TRUE. Then the first variable of the dta-file will hold the rownames of the resulting data.frame.

Reading dta-files of older and newer versions than 13 was introduced with version 0.8.

References