Performance

[jmboehm]

Many of the macros are generating code that is not type stable, and is therefore horrendously slow. This affects in particular

• egen/erep (fixed with first commit)
• generate/replace <-- should use Tables.namedtupeiterator and a barrier function
• reshape maybe more?

[codecov[bot]]

Codecov Report

Merging #11 into master will increase coverage by 1.05%. The diff coverage is 92.96%.

@@            Coverage Diff             @@
##           master      #11      +/-   ##
==========================================
+ Coverage   76.64%   77.69%   +1.05%     
==========================================
  Files          17       18       +1     
  Lines         608      677      +69     
==========================================
+ Hits          466      526      +60     
- Misses        142      151       +9     

Impacted Files	Coverage Δ
src/Douglass.jl	100.00% <ø> (ø)
src/commands/duplicates.jl	100.00% <ø> (ø)
src/helper.jl	88.23% <86.20%> (-7.22%)	:arrow_down:
src/commands/erep.jl	93.33% <93.33%> (ø)
src/commands/egen.jl	95.91% <96.87%> (-1.27%)	:arrow_down:
src/commands/drop.jl	100.00% <100.00%> (ø)
src/commands/generate.jl	91.42% <100.00%> (ø)
src/commands/keep.jl	100.00% <100.00%> (ø)
src/commands/replace.jl	91.89% <100.00%> (ø)
src/commands/reshape.jl	65.00% <100.00%> (+0.89%)	:arrow_up:
... and 4 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 09cd0bb...b4b5fa8. Read the comment docs.

[jmboehm]

Merging this sooner rather than later. reshape_long is still painfully slow, which I think is coming from the fact that DataFrames.stack is slow (probably not type stable).

Some bug fixes too.

Also, this is changing the version to 0.0.1, in preparation for registration.