Symbol column indexing

.travis.yml

@@ -1,11 +1,15 @@
 language: julia
+
 os:
   - linux
+
 julia:
   - 0.7
   - 1.0
+
 notifications:
   email: false
+
 sudo: false
 
 after_success:

NEWS.md

@@ -1,3 +1,47 @@
+### 0.14.0
+
+* Symbol column indexing. ([#377])
+  And the String indexing is deprecated.
+
+  E.g.
+
+  ```julia
+  using MarketData
+  ohlc[:Close]  # and cl["Close"] is deprecated
+  ```
+
+* `Base.getproperty` support. ([#377])
+  All the columns can be accessed via the form `ta.column_name`.
+
+  E.g.
+
+  ```julia
+  using MarketData
+  ohlc.Open
+  ```
+
+  The original `TimeArray` fields getters is available as functions.
+
+    * `timestamp(::TimeArray)`
+    * `values(::TimeArray)`
+    * `colnames(::TimeArray)`
+    * `meta(::TimeArray)`
+
+  ```julia
+  ohlc.values  # this is unavailable due to Base.getproperty support
+  values(ohlc)  # change to this
+  ```
+
+[#377]: https://github.com/JuliaStats/TimeSeries.jl/pull/377
+
+
+### 0.13.0
+
+* Julia v0.7/1.0 support. ([#370])
+
+[#370]: https://github.com/JuliaStats/TimeSeries.jl/pull/370
+
+
 ### 0.12.0
 
 * Support `Base.copy(::TimeArray)`. ([#352])

docs/src/apply.md

@@ -135,6 +135,7 @@ In our moving average example, we would pass arguments this way:
 ```@repl
 using TimeSeries
 using MarketData
+using Statistics
 moving(mean, cl, 10)
 ```
 

docs/src/combine.md

@@ -85,6 +85,7 @@ We can also supply the function that chooses the timestamp and the
 function that determines the corresponding value independently:
 
 ```@repl collapse
+using Statistics
 collapse(cl, month, last, mean)
 ```
 
@@ -104,8 +105,8 @@ For example:
 
 ```@repl
 using TimeSeries
-a = TimeArray([Date(2015, 10, 01), Date(2015, 11, 01)], [15, 16], ["Number"])
-b = TimeArray([Date(2015, 12, 01)], [17], ["Number"])
+a = TimeArray([Date(2015, 10, 01), Date(2015, 11, 01)], [15, 16])
+b = TimeArray([Date(2015, 12, 01)], [17])
 [a; b]
 ```
 
@@ -124,6 +125,7 @@ series, here by one year:
 
 ```@repl
 using TimeSeries
-a = TimeArray([Date(2015, 10, 01), Date(2015, 11, 01)], [15, 16], ["Number"])
-map((timestamp, values) -> (timestamp + Dates.Year(1), values), a)
+using Dates
+ta = TimeArray([Date(2015, 10, 01), Date(2015, 11, 01)], [15, 16])
+map((timestamp, values) -> (timestamp + Year(1), values), ta)
 ```

docs/src/dotfile.md

@@ -7,7 +7,9 @@ and only controls how values are printed to REPL.
 Here is an handy way to edit it:
 
 ```julia
-julia> edit(Pkg.dir("TimeSeries", "src", ".timeseriesrc.jl"))
+julia> using TimeSeries
+
+julia> edit(joinpath(dirname(pathof(TimeSeries)), ".timeseriesrc.jl"))
 ```
 
 ## `DECIMALS`

docs/src/getting_started.md

@@ -22,7 +22,8 @@ the following code block:
 using TimeSeries
 ```
 
-```@example dummy
-dates  = collect(Date(1999,1,1):Date(2000,12,31))
-mytime = TimeArray(dates, rand(length(dates)))
+```@repl dummy
+using Dates
+dates = Date(2018, 1, 1):Day(1):Date(2018, 12, 31)
+ta = TimeArray(dates, rand(length(dates)))
 ```

docs/src/indexing.md

@@ -2,67 +2,88 @@
 
 Indexing out a time series is done with common bracketing semantics.
 
-## Integers
+## Row indexing
+
+### Integer
 
 | Example      | Description                           | Indexing value                 |
 |--------------|---------------------------------------|--------------------------------|
 | `[1]`        | First row of data only                | single integer                 |
 | `[1:3]`      | First through third row only          | integer range                  |
 | `[1:2:10]`   | Odd row between first to tenth row    | integer range with step        |
 | `[[1:3; 8]]` | First through third row and eight row | integer range & single integer |
+| `[end]`      | Last row                              |                                |
 
 Examples in REPL:
 
-```@repl
+```@setup int-indexing
 using MarketData
+```
+
+```@repl int-indexing
 ohlc[1]
 ohlc[1:3]
 ohlc[1:2:10]
 ohlc[[1:3;8]]
+ohlc[end]
 ```
 
-## Strings
+### Date and DateTime
 
-| Example             | Description                          | Indexing value   |
-|---------------------|--------------------------------------|------------------|
-| `["Open"]`          | The column named "Open"              | single string    |
-| `["Open", "Close"]` | The columns named "Open" and "Close" | multiple strings |
+| Example                                      | Description                                | Indexing value |
+|----------------------------------------------|--------------------------------------------|----------------|
+| `[Date(2000, 1, 3)]`                         | The row containing Jan 3, 2000 timestamp   | single Date    |
+| `[[Date(2000, 1, 3), Date(2000, 2, 4)]]`     | The rows containing Jan 3 & Jan 4, 2000    | multiple Dates |
+| `[Date(2000, 1, 3):Day(1):Date(2000, 2, 4)]` | The rows between Jan 1, 2000 & Feb 1, 2000 | range of Date  |
 
 Examples in REPL:
 
-```@repl
+```@setup date-indexing
 using MarketData
-ohlc["Open"]
-ohlc["Open", "Close"]
+using Dates
+```
+
+```@repl date-indexing
+ohlc[Date(2000, 1, 3)]
+ohlc[Date(2000, 1, 3):Day(1):Date(2000, 2, 4)]
 ```
 
-# Date and DateTime
+## Column indexing
 
-| Example                                  | Description                                | Indexing value |
-|------------------------------------------|--------------------------------------------|----------------|
-| `[Date(2000, 1, 3)]`                     | The row containing Jan 3, 2000 timestamp   | single Date    |
-| `[[Date(2000, 1, 3), Date(2000, 2, 4)]]` | The rows containing Jan 3 & Jan 4, 2000    | multiple Dates |
-| `[Date(2000, 1, 3):Date(2000, 2, 4)]`    | The rows between Jan 1, 2000 & Feb 1, 2000 | range of Date  |
+### Symbol
+
+| Example           | Description                            | Indexing value   |
+|-------------------|----------------------------------------|------------------|
+| `[:Open]`         | The column named `:Open`               | single symbol    |
+| `[:Open, :Close]` | The columns named `:Open` and `:Close` | multiple symbols |
 
 Examples in REPL:
 
-```@repl
+```@setup symbol-indexing
 using MarketData
-ohlc[Date(2000, 1, 3)]
-ohlc[Date(2000, 1, 3):Date(2000, 2, 4)]
+using Dates
+```
+
+```@repl symbol-indexing
+ohlc[:Open]
+ohlc[:Open, :Close]
 ```
 
 ## Mixed approach
 
-| Example                      | Description                   | Indexing value                |
-|------------------------------|-------------------------------|-------------------------------|
-| `["Open"][1:3]`              | "Open" column & first 3 rows  | single string & integer range |
-| `["Open"][Date(2000, 1, 3)]` | "Open" column and Jan 3, 2000 | single string & Date          |
+| Example                     | Description                    | Indexing value                |
+|-----------------------------|--------------------------------|-------------------------------|
+| `[:Open][1:3]`              | `:Open` column & first 3 rows  | single symbol & integer range |
+| `[:Open][Date(2000, 1, 3)]` | `:Open` column and Jan 3, 2000 | single symbol & Date          |
 
 Examples in REPL:
 
-```@repl
+```@setup mixed-indexing
 using MarketData
-ohlc["Open"][1:3]
-ohlc["Open"][Date(2000, 1, 3)]
+using Dates
+```
+
+```@repl mixed-indexing
+ohlc[:Open][1:3]
+ohlc[:Open][Date(2000, 1, 3)]
 ```

docs/src/modify.md

@@ -1,4 +1,4 @@
-# Modify existing `TimeArrays`
+# Modify existing `TimeArray`s
 
 Since `TimeArrays` are immutable, they cannot be altered or changed
 in-place. In practical application, an existing `TimeArray` might need to
@@ -18,7 +18,7 @@ using TimeSeries
 using MarketData
 update(cl, Date(2002,1,1), 111.11)
 update(cl, Date(2002,1,1), [111.11])
-update(ohlc, Date(2002,1,1), [111.11, 222.22, 333.33, 444.44])
+update(ohlc, Date(2002,1,1), [111.11 222.22 333.33 444.44])
 ```
 
 ## `rename`
@@ -28,7 +28,7 @@ The `rename` method allows the column name(s) to be changed:
 ```@repl
 using TimeSeries
 using MarketData
-rename(cl, "New Close")
-rename(cl, ["New Close"])
-rename(ohlc, ["New Open", "New High", "New Low", "New Close"])
+rename(cl, :Close′)
+rename(cl, [:Close′])
+rename(ohlc, [:Open′, :High′, :Low′, :Close′])
 ```

docs/src/plotting.md

@@ -24,7 +24,7 @@ More sophisticated plots can be created by using keyword attributes and
 subsets:
 
 ```julia
-plot(MarketData.ohlc["Low"], seriestype = :scatter, markersize = 3, color = :red, markeralpha = 0.4, grid = true)
+plot(MarketData.ohlc[:Low], seriestype = :scatter, markersize = 3, color = :red, markeralpha = 0.4, grid = true)
 ```
 
 ![image](images/complexplot.svg)

docs/src/readwrite.md

@@ -4,7 +4,7 @@ Reading/writing a `csv` file into a `TimeArray` object is supported.
 
 ## `readtimearray`
 
-The `readtimearray` method is a wrapper for the `Base.readcsv` method
+The `readtimearray` method is a wrapper for the `DelimitedFiles.readdlm` method
 that returns a `TimeArray`.
 
 ```julia
@@ -16,7 +16,7 @@ of the `csv` file that you wish to parse into a `TimeArray` object.
 Optionally, you can add a value to the `meta` field.
 
 More generally, this function accepts arbitrary delimiters with `delim`,
-just like `Base.readcsv`.
+just like `DelimitedFiles.readdlm`.
 
 For `DateTime` data that has odd formatting, a `format` argument is
 provided where users can pass the format of their data.

docs/src/split.md

@@ -1,4 +1,4 @@
-# Splitting by time constraint or when condition is true
+# Splitting by conditions
 
 Specific methods for segmenting on time ranges or if condition is met is
 supported with the following methods.
@@ -65,21 +65,21 @@ The `findwhen` method test a condition and returns a vector of `Date` or
 using TimeSeries
 using MarketData
 
-green = findwhen(ohlc["Close"] .> ohlc["Open"]);
+green = findwhen(ohlc[:Close] .> ohlc[:Open]);
 typeof(green)
 ohlc[green]
 ```
 
-## `find`
+## `findall`
 
-The `find` method tests a condition and returns a vector of `Int`
+The `findall` method tests a condition and returns a vector of `Int`
 representing the row in the array where the condition is `true`:
 
 ```@repl
 using TimeSeries
 using MarketData
 
-red = find(ohlc["Close"] .< ohlc["Open"]);
+red = findall(ohlc[:Close] .< ohlc[:Open]);
 typeof(red)
 ohlc[red]
 ```

docs/src/timearray.md

@@ -4,10 +4,10 @@ The `TimeArray` time series type is defined here
 (with inner constructor code removed for readability):
 
 ```julia
-struct TimeArray{T,N,D<:TimeType,A<:AbstractArray} <: AbstractTimeSeries
+struct TimeArray{T,N,D<:TimeType,A<:AbstractArray{T,N}} <: AbstractTimeSeries{T,N,D}
     timestamp::Vector{D}
     values::A # some kind of AbstractArray{T,N}
-    colnames::Vector{String}
+    colnames::Vector{Symbol}
     meta::Any
 
     # inner constructor code enforcing invariants
@@ -37,7 +37,7 @@ be of the same type.
 
 ## `colnames`
 
-The `colnames` field is a vector of type `String` and contains the
+The `colnames` field is a vector of `Symbol` and contains the
 names of the columns for each column in the `values` field. The length
 of this vector must match the column count of the `values` array, or the
 constructor will fail. Since `TimeArrays` are indexable on column names,
@@ -48,8 +48,18 @@ where `n` enumerates from 1.
 ## `meta`
 
 The `meta` field defaults to holding `nothing`, which is represented by
-type `Void`. This default is designed to allow programmers to ignore
+type `Nothing`. This default is designed to allow programmers to ignore
 this field. For those who wish to utilize this field, `meta` can hold
 common types such as `String` or more elaborate user-defined types. One
 might want to assign a name to an object that is immutable versus
 relying on variable bindings outside of the object's type fields.
+
+## Fields getter functions
+
+There are four field getter functions exported.
+They are named as same as the field names.
+
+- `timestamp(ta::TimeArray)`
+- `values(ta::TimeArray)`
+- `colnames(ta::TimeArray)`
+- `meta(ta::TimeArray)`