m61r 0.1.0
Major New Features
- Enhanced “Pipeline” Interface: The
m61r object now allows for a more fluid chaining of
“pipe-like” operations.
- Support for Scoped Operations:
- Introduced
.SD() and .select_cols() for
advanced column manipulation within the object’s environment.
- Added the
across() function, allowing a function to be
applied to multiple columns simultaneously within mutate()
and summarise().
- Temporal Orchestration:
- Added the
$explode() method to explode
list-columns, typically used after creating temporal sequences.
- Added
$join_asof() for proximity-based temporal joins
(backward/forward), essential for asynchronous time-series.
- Conditional Logic: Implemented
case_when_ to allow complex vectorised syntax within the
m61r object.
- I/O: Implemented
write.csv to allow
exporting data.frame from the m61r object.
API Improvements
- summarise_: Complete code refactoring of name
management. Named expressions (e.g.,
avg = ~mean(x)) and
list expressions (e.g., ~across(...)) now correctly handle
suffixes and prevent NA column names.
- mutate_: Optimized internal evaluation for better
memory management on large data frames.
- Joins: Improved stability of join functions
(
inner, left, right,
full, semi, anti) across mixed
column types.
Documentation
- Created two new detailed vignettes:
m61r_advanced: Complex manipulation and functional
programming.
m61r_temporal: Specific management of time-series
data.