Julia 0.7 Introduces First-Class Support for Statistical Missing Values
The upcoming Julia 0.7 release introduces first-class support for statistical missing values, a feature essential for data management and analysis. Unlike general-purpose languages that rely on Nullable or Option types, Julia 0.7 implements a new `missing` object, succeeding the `NA` value from the DataArrays package. This design draws inspiration from SQL's NULL and R's NA, facilitating easier interoperability with these systems. The new implementation allows `missing` to combine with any type, ensuring safety by default through propagation or error throwing, while maintaining high performance using standard Array types instead of specialized DataArrays. This update addresses limitations in previous approaches, such as sentinel values used by R and Pandas, which restrict missing value representation to specific domains. By leveraging improvements in the Julia compiler, the `missing` object offers efficient, generic handling of incomplete data without requiring special coding tricks. This framework is already utilized by DataFrames version 0.11, marking a significant advancement for scientific computing in Julia.
Wire timeline
Julia 0.7 Introduces First-Class Support for Statistical Missing Values
The upcoming Julia 0.7 release introduces first-class support for statistical missing values, a feature essential for data management and analysis. Unlike general-purpose languages that rely on Nullable or Option types, Julia 0.7 implements a new `missing` object, succeeding the `NA` value from the DataArrays package. This design draws inspiration from SQL's NULL and R's NA, facilitating easier interoperability with these systems. The new implementation allows `missing` to combine with any type, ensuring safety by default through propagation or error throwing, while maintaining high performance using standard Array types instead of specialized DataArrays. This update addresses limitations in previous approaches, such as sentinel values used by R and Pandas, which restrict missing value representation to specific domains. By leveraging improvements in the Julia compiler, the `missing` object offers efficient, generic handling of incomplete data without requiring special coding tricks. This framework is already utilized by DataFrames version 0.11, marking a significant advancement for scientific computing in Julia.
JuliaLang - The Julia programming language