A Package to read and write Stata .dta files. This may be the first Julia DataFrame to Stata conversion program. Despite the excellent .dta file reader in ReadStat.jl package, I needed to write another .dta file reader because ReadStat.jl was not able to handle very large files well and all formats such as value labels and date and datatime formats are lost during conversion. It is written entirely in Julia.

StataIO.jl provides the following functionalities:

• When importing, it can break large files into small chunks and so can handle very large files gracefully in memory strapped situations. Use chunks option to specify the number of chunks to use (default = 10).

• All variable labels will be imported using TableMetadataTools.jl.

• All variables with value labels will be imported into CategoricalArrays. If keep_original = false is set, only the value labels will be imported and the original numeric values will NOT be kept. If you want to keep the original values, use keep_original = true. Then, original Stata numeric values will be prepended to the value labels followed by a : (default: false).

• All int or long variables with %d or %td formats will be imported as Julia dates using the Dates package.

• All float or double variables with %tc or %tC formats will be imported as Julia DateTime variables.

• When exporting, Julia Dates and DateTimes will be automatically converted to Stata with the formats %tdNN-DD-CCYY for dates and %tc for datetimes.

• CategoricalArrays will be exported into byte or int types with all value labels.

• Any variables with data types that are not supported by Stata (e.g., Int64) will NOT be exported.

] add StataIO

df = read_stata(Statafilename::String; chunks::Int = 10, keep_original = false)

write_stata(Statafilename::String, df::AbstractDataFrame)