Skip to content

Commit

Permalink
fixing gammaCKpar! redef
Browse files Browse the repository at this point in the history
  • Loading branch information
jw2249a committed Jan 4, 2024
1 parent 550c60e commit 369160d
Show file tree
Hide file tree
Showing 2 changed files with 7 additions and 6 deletions.
11 changes: 6 additions & 5 deletions src/benchmark.jl
Original file line number Diff line number Diff line change
Expand Up @@ -10,11 +10,12 @@ b_fil="../../rstudio/test_merge/data/test_b.csv"

#varnames=["FIRST_NAME"]
varnames=["FIRST_NAME", "MIDDLE_NAME", "LAST_NAME", "STREET_NAME"]
match_type=["string_partial", "string", "exact", "string_partial"]
#match_type=["string_partial", "string", "exact", "string_partial"]
match_type=["string_partial", "string_partial", "string_partial", "string_partial"]
#varnames=["FIRST_NAME", "MIDDLE_NAME", "LAST_NAME", "STREET_NAME", "STATE"]
#[100,200,500,1_000,2_000,4_000, 5_000, 10_000,20_000, 40_000, 50_000,100_000,1_000_000]
N2=10_000
N1_N=[1_000,10_000,50_000,100_000,500_000,1_000_000]
N1_N=[1_000,10_000,50_000,100_000,500_000,750_000,1_000_000]
println("## $(length(varnames)) vars")
for N1 in N1_N

Expand All @@ -36,7 +37,7 @@ for N1 in N1_N
println(center_in_line("(FUZZY🧸🐈🦭)", pad_char='-'))
@btime FastLink.fastLink($dfA,$dfB,$varnames,$match_type,fuzzy=true)
println("")
println(center_in_line("(NOT FUZZY🐬🐳)", pad_char='-'))
@btime FastLink.fastLink($dfA,$dfB,$varnames,$match_type,fuzzy=false)
println("")
# println(center_in_line("(NOT FUZZY🐬🐳)", pad_char='-'))
# @btime FastLink.fastLink($dfA,$dfB,$varnames,$match_type,fuzzy=false)
# println("")
end
2 changes: 1 addition & 1 deletion src/gammaCKfuzzy.jl
Original file line number Diff line number Diff line change
Expand Up @@ -234,7 +234,7 @@ https://tech.popdata.org/speeding-up-Jaro-Winkler-with-rust-and-bitwise-operatio
- `upper::Bool=true`: Whether input string is uppercase.
- `w`: Winkler weight for jw string distance.
"""
function gammaCKfuzzy!(vecA::PooledVector,vecB::PooledVector, results::SubArray, array_2Dindex::Function, dims::Tuple; cut_a=0.92, upper=true,w=0.1)
function gammaCK2fuzzy!(vecA::PooledVector,vecB::PooledVector, results::SubArray, array_2Dindex::Function, dims::Tuple; cut_a=0.92, upper=true,w=0.1)

# functions that update the results view
function update_results!(a_ids::Vector{UInt32},
Expand Down

0 comments on commit 369160d

Please sign in to comment.