updated to dataframes v1

Project.toml

@@ -1,7 +1,7 @@
 name = "JDF"
 uuid = "babc3d20-cd49-4f60-a736-a8f9c08892d3"
 authors = ["Dai ZJ <zhuojia.dai@gmail.com>"]
-version = "0.4.1"
+version = "0.4.2"
 
 [deps]
 Blosc = "a74b3585-a348-5f62-a45c-50e91977d574"
@@ -21,7 +21,7 @@ Blosc = "0.5, 0.6, 0.7"
 BufferedStreams = "1.0"
 CategoricalArrays = "0.5, 0.6, 0.7, 0.8, 0.9"
 DataAPI = "1"
-Missings = "0.4"
+Missings = "1"
 PooledArrays = "1"
 StatsBase = "0.32, 0.33"
 Tables = "1"

README.jl

@@ -6,15 +6,15 @@ a = dataset("datasets", "iris");
 first(a, 2)
 
 
-@time jdffile = savejdf("iris.jdf", a)
-@time a2 = loadjdf("iris.jdf")
+@time jdffile = JDF.save("iris.jdf", a)
+@time a2 = DataFrame(JDF.load("iris.jdf"))
 
 
 all(names(a2) .== names(a)) # true
 all(skipmissing([all(a2[!,name] .== Array(a[!,name])) for name in names(a2)])) #true
 
 
-a2_selected = loadjdf("iris.jdf", cols = [:Species, :SepalLength, :PetalWidth])
+a2_selected = DataFrame(JDF.load("iris.jdf", cols = [:Species, :SepalLength, :PetalWidth]))
 
 
 jdf"path/to/JDF.jdf"

README.jmd

@@ -69,25 +69,16 @@ path_to_JDF = "path/to/JDF.jdf"
 JDFFile(path_to_JDF)
 ```
 
-#### Using `df[rows, cols]` syntax
-You can load arbitrary `rows` and `cols` using the `df[rows, cols]` syntax. However, some of these operations are not yet optimized and hence may not be efficient.
+#### Using `df[col::Symbol]` syntax
+You can load arbitrary `col` using the `df[col]` syntax. However, some of these operations are not
+yet optimized and hence may not be efficient.
 
 ```julia
 afile = JDFFile("iris.jdf")
 
-afile[!, :Species] # load Species column
-afile[!, [:Species, :PetalLength]] # load Species and PetalLength column
-
-afile[:, :Species] # load Species column
-afile[:, [:Species, :PetalLength]] # load Species and PetalLength column
-
-@view(afile[!, :Species]) # load Species column
-@view(afile[!, [:Species, :PetalLength]]) # load Species and PetalLength column
+afile[:Species] # load Species column
 ```
 
-In fact most syntax for `a[rows, cols]` will work **except** for assignments i.e. `a[!, cols] = something` will **not** work.
-
-This was developed to make it possible for [JLBoost.jl](https://github.com/xiaodaigh/JLBoost.jl) to fit models without loading the whole data into memory, and so the functionalities is kept to a minimum for now.
 
 #### JDFFile is Table.jl columm-accessible
 
@@ -132,28 +123,17 @@ end
 ```
 
 #### Metadata Names & Size from disk
-You can obtain the column names and size (`nrow` and `ncol`) of a JDF, for
+You can obtain the column names and number of columns `ncol` of a JDF, for
 example:
 
 
 ```julia
 using JDF, DataFrames
 df = DataFrame(a = 1:3, b = 1:3)
-savejdf(df, "plsdel.jdf")
-
+JDF.save(df, "plsdel.jdf")
 
 names(jdf"plsdel.jdf") # [:a, :b]
 
-nrow(jdf"plsdel.jdf") # 3
-
-ncol(jdf"plsdel.jdf") # 2
-
-size(jdf"plsdel.jdf") # (3, 2)
-
-size(jdf"plsdel.jdf", 1) # 2
-
-size(jdf"plsdel.jdf", 2) # 3
-
 # clean up
 rm("plsdel.jdf", force = true, recursive = true)
 ```