CHANGELOG.md

v0.4

• Technically breaking: A basically unknown functionality of pre-0.3 versions, using arbitrary column specifiers on the left-hand side, was removed. Instead, the behavior of using {} and {n} expressions on the left hand side was assimilated to the existing shortcut string syntax. {} or {1} refers to the name of the first column used, {2} to the second, etc. This allows to use transformation expressions on the used column names, such as uppercase({}) or split({2})[1].

v0.3.3

• Fixed bug when column symbols were used inside braces.

v0.3.2

• Added @proprow, @eachindex and @groupindices special function macros, which required compat to be raised to DataFrames v1.4.

v0.3.0

• Breaking: The $() interpolation syntax is replaced by {} for single columns (or broadcasted multi-columns)
• Added {{}} for referring to multiple columns as a tuple.
• Breaking: No more flag macros, replaced by explit @byrow, @bycol, @passmissing, @astable.
• Breaking: All(), Between() and Not() have to be interpolated with {} and can't be used standalone anymore.

v0.2.1

• Added ability to use multi-column expressions with All(), Between(), Not(), as well as any other multi-column identifier such as $Real for all columns of eltype Real.
• Added shortcut string option to specify renaming structures, e.g. @transform(df, "sqrt_of_{}" = sqrt(All())).

v0.2

• Added @subset argument for @transform! and @select! that performs a mutating transform on a subset of rows of a DataFrame.

v0.1.2

• Fixed macro hygiene and enabled Julia 1.0.

v0.1.1

• Added tuple destructuring syntax for the @t flag macro. #14

v0.1

• Initial release.