Merge pull request #51 from ExpandingMan/tables

switch from DataStreams to Tables

Author: aviks

.travis.yml

@@ -1,12 +1,28 @@
+
+sudo: required
+
 language: julia
+
+services:
+  - docker
+
 os:
+  - osx
   - linux
+
 julia:
-    - 0.7
-    - 1.0
+  - 1.0
+  - 1.3
+  - nightly
+
+matrix:
+  allow_failures:
+  - julia: nightly
+
 notifications:
   email: false
-# uncomment the following lines to override the default test script
-#script:
-#  - if [[ -a .git/shallow ]]; then git fetch --unshallow; fi
-#  - julia --check-bounds=yes -e 'Pkg.clone(pwd()); Pkg.build("JDBC"); Pkg.test("JDBC"; coverage=true)'
+
+after_success:
+  - julia -e 'import Pkg, JDBC; cd(joinpath(dirname(pathof(JDBC)),"..")); Pkg.add("Coverage"); using Coverage; Codecov.submit(process_folder())'
+  - julia -e 'import Pkg; Pkg.add("Documenter")'
+  - julia -e 'import JDBC; cd(joinpath(dirname(pathof(JDBC)),"..")); include(joinpath("docs", "make.jl"))'

Project.toml

@@ -1,14 +1,13 @@
 name = "JDBC"
 uuid = "6042db11-3c3d-5e84-8dba-9cbf74c9ba48"
-version = "0.4.1"
+version = "0.4.2"
 
 [deps]
-DataStreams = "9a8bc11e-79be-5b39-94d7-1ccc349a1a85"
 Dates = "ade2ca70-3891-5945-98fb-dc099432e06a"
 JavaCall = "494afd89-becb-516b-aafa-70d2670c0337"
+Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
 
 [compat]
-DataFrames = "< 0.18.0"
 JavaCall = "≥ 0.7.0"
 julia = "≥ 0.7.0"
 

README.md

@@ -6,7 +6,7 @@
 This package enables the use of Java JDBC drivers to access databases from within Julia. It uses the [JavaCall.jl](https://github.com/aviks/JavaCall.jl) package to call into Java in order to use the JDBC drivers. 
 
 The API provided by this package consists essentially of two components: a "direct" (i.e. minimally wrapped) interface directly to Java JDBC and a minimal
-Julian interface with support for [DataStreams.jl](https://github.com/JuliaData/DataStreams.jl).
+Julian interface with support for [Tables.jl](https://github.com/JuliaData/Tables.jl).
 
 This package currently supports only Julia v0.6 and later.
 
@@ -121,13 +121,13 @@ end
 close(csr)  # closes Connection, can be called on Connection or Cursor
 ```
 
-#### `DataStreams` Interface and Creating `DataFrame`s
+#### `Tables` Interface and Creating `DataFrame`s
 
-JDBC includes a [DataStreams](https://github.com/JuliaData/DataStreams.jl) interface.  A DataStreams `Source` object can be created from a `JDBC.Cursor` or a
-`JDBCRowIterator` simply by doing e.g. `JDBC.Source(csr)`.  This object implements the DataStreams `Data.Source` interface.  It can be useful for retrieving metadata
-with `Data.schema`.
+JDBC includes a [Tables](https://github.com/JuliaData/Tables.jl) interface.  A Tables
+`Source` object can be created from a `JDBC.Cursor` or a `JDBCRowIterator` simply by doing
+e.g. `JDBC.Source(csr)`.  It can be useful for retrieving metadata with `Tables.schema`.
 
-This is also useful for loading data from a database into an object that implements the DataStreams `Data.Sink` interface such as a `DataFrame`.  For this we
+This is also useful for loading data from a database into another object that implements the Tables interface.  For this we
 provide also the convenient `JDBC.load` function.
 
 For example, you can do

REQUIRE

@@ -1,7 +1,7 @@
 environment:
   matrix:
-  - julia_version: 0.7
-  - julia_version: 1
+  - julia_version: 1.0
+  - julia_version: 1.3
   - julia_version: nightly
 
 platform:
@@ -12,8 +12,7 @@ platform:
 # # (tests will run but not make your overall status red)
 matrix:
   allow_failures:
-  - julia_version: nightly
-  - platform: x86
+    - julia_version: nightly
 
 branches:
   only:

appveyor.yml

@@ -640,7 +640,7 @@ export getTableMetaData, JDBCRowIterator
 
 
 include("interface.jl")
-include("datastreams.jl")
+include("tables.jl")
 
 
 end # module

src/JDBC.jl

@@ -1,4 +1,4 @@
-using DataStreams
+using Tables
 
 const column_types = Dict(
                           JDBC_COLTYPE_ARRAY=>Array,
@@ -57,41 +57,34 @@ colname(s::Source, col::Int) = getColumnName(s.md, col)
 ncols(s::Source) = getColumnCount(s.md)
 
 coltypes(s::Source) = Type[coltype(s, i) for i ∈ 1:ncols(s)]
-colnames(s::Source) = String[colname(s, i) for i ∈ 1:ncols(s)]
+colnames(s::Source) = Symbol[Symbol(colname(s, i)) for i ∈ 1:ncols(s)]
 
-# WARNING: this does not seem to actually work
-Data.reset!(s::Source) = beforeFirst!(s.rs)
+Tables.istable(::Type{<:Source}) = true
+Tables.rowaccess(::Type{<:Source}) = true
+Tables.rows(s::Source) = s
+Tables.schema(s::Source) = Tables.Schema(colnames(s), coltypes(s))
 
-Data.isdone(s::Source, row::Int, col::Int)  = isdone(s.rs)
-
-Data.schema(s::Source) = Data.Schema(coltypes(s), colnames(s), missing)
-
-Data.accesspattern(s::Source) = Data.Sequential
-
-Data.streamtype(::Type{Source}, ::Type{Data.Field}) = true
-Data.streamtype(::Type{Source}, ::Type{Data.Column}) = false
+Base.IteratorSize(::Type{<:Source}) = Base.SizeUnknown()
+Base.eltype(s::Source) = namedtupletype(Tables.schema(s))
+namedtupletype(::Tables.Schema{names, types}) where {names, types} = NamedTuple{names, types}
+namedtupletype(s::Source) = namedtupletype(Tables.schema(s))
 
 # TODO currently jdbc_get_method is very inefficient
 pullfield(s::Source, col::Int) = jdbc_get_method(getColumnType(s.md, col))(s.rs, col)
 
-# does not store current row number as a persistent state
-function Data.streamfrom(s::Source, ::Type{Data.Field}, ::Type{T}, row::Int, col::Int) where T
-    convert(T, pullfield(s, col))::T
-end
-function Data.streamfrom(s::Source, ::Type{Data.Field}, ::Type{Union{T, Missing}},
-                         row::Int, col::Int) where T
-    o = pullfield(s, col)
-    if wasNull(s.rs)
-        return missing
-    end
-    convert(T, o)::T
+jdbcconvert(::Type{T}, s, x) where {T} = convert(T, x)
+jdbcconvert(::Type{Union{T, Missing}}, s, x) where {T} = wasNull(s.rs) ? missing : convert(T, x)
+
+function Base.iterate(s::Source, NT::Type{NamedTuple{names, types}}=namedtupletype(s)) where {names, types}
+    isdone(s.rs) && return nothing
+    NT(jdbcconvert(fieldtype(types, i), s, pullfield(s, i)) for i ∈ 1:fieldcount(types)), NT
 end
 
-load(::Type{T}, s::Source) where {T} = Data.close!(Data.stream!(s, T))
+load(::Type{T}, s::Source) where {T} = T(Tables.materializer(T)(s))
 load(::Type{T}, rs::JResultSet) where {T} = load(T, Source(rs))
 load(::Type{T}, stmt::JStatement, query::AbstractString) where {T} = load(T, Source(stmt, query))
 load(::Type{T}, csr::Union{JDBC.Cursor,JDBCRowIterator}) where {T} = load(T, Source(csr))
-function load(::Type{T}, csr::Cursor, q::AbstractString) where T
+function load(::Type{T}, csr::Cursor, q::AbstractString) where {T}
     execute!(csr, q)
     load(T, csr)
 end

src/datastreams.jl src/tables.jlsrc/datastreams.jl renamed to src/tables.jl

@@ -2,7 +2,6 @@
 using JavaCall
 using JDBC
 using DataFrames
-using DataStreams
 using Test
 using Dates
 import Pkg