以 Alpha 101 为例：量化因子的多范式计算方法

@state
def alpha1TS(close){
    return mimax(pow(iif(ratios(close) - 1 < 0, mstd(ratios(close) - 1, 20),close), 2.0), 5)
}
def alpha1Panel(close){
    return rowRank(X=alpha1TS(close), percent=true) - 0.5 //rowRank：在每一个时间截面对标的排名
}
raw_data_alpha1 = select * from loadTable("dfs://k_minute_level","k_minute") where date(tradetime) between 2020.02.01 : 2020.02.28 
input_panel_alpha1 = exec close from raw_data_alpha1 pivot by tradetime, securityid 
alpha1_panel_res = alpha1Panel(input_panel_alpha1)

def prepareDataForDDBPanel(raw_data, start_time, end_time){
    t = select tradetime,securityid, vwap,vol,open from raw_data where date(tradetime) between start_time : end_time
    return dict(`vwap`open`vol, panel(t.tradetime, t.securityid, [t.vwap, t.open, t.vol]))
}
@state
def alpha98Panel(vwap, open, vol){
    return rowRank(X=mavg(mcorr(vwap, msum(mavg(vol, 5), 26), 5), 1..7),percent=true) - rowRank(X=mavg(mrank(9 - mimin(mcorr(rowRank(X=open,percent=true), rowRank(X=mavg(vol, 15),percent=true), 21), 9), true, 7), 1..8),percent=true)
}
raw_data_alpha98 = loadTable("dfs://k_minute_level","k_minute")
start_time = 2020.02.01
end_time = 2020.02.28
input_panel_alpha98 = prepareDataForDDBPanel(raw_data_alpha98, start_time, end_time)
alpha98DDBPanel = alpha98Panel(input_panel_alpha98.vwap, input_panel_alpha98.open, input_panel_alpha98.vol)

各矩阵内部运用了 rowRank 函数进行横向截面运算，以及 m 系列函数进行垂直滑动窗口计算，各矩阵之间也进行了二元运算。最后仅用一行代码就解决了多维度的复杂的嵌套计算逻辑。

其实，面板模式在 Python、Matlab 等一些专门用于数据分析的工具中也都可以实现，但其计算过程主要基于内存运算。也就是说，随着数据频率上升，数据量增加，面板模式对分钟频、快照、逐笔等中高频数据的处理会变得吃力——为此，掌握 DolphinDB 的 SQL 写法就非常重要。

DolphinDB 对 SQL 的支持，搭配上分布式系统，体现出了以下优势：

def sum_diff(x, y){
    return (x-y)/(x+y)
}


ema(1000 * sum_diff(ema(price, 20), ema(price, 40)),10) -  ema(1000 * sum_diff(ema(price, 20), ema(price, 40)), 20)

我们可以直接在 SQL 中通过 context by 子句按股票代码分组，然后在 select 中调用 factorDoubleEMA 因子计算函数，来计算每个股票在一段时间里的因子值。

@state
def factorDoubleEMA(price){
    ema_2 = ema(price, 2) //ema：在给定长度（以元素个数衡量）的滑动窗口内，计算 X 的指数移动平均(Exponential Moving Average)。
    ema_4 = ema(price, 4)
    sum_diff_1000 = 1000 * sum_diff(ema_2, ema_4)
    return ema(sum_diff_1000, 2) - ema(sum_diff_1000, 3)
}
res = select tradetime, securityid, `doubleEMA as factorname, factorDoubleEMA(close) as val from loadTable("dfs://k_minute_level","k_minute") where date(tradetime) between 2020.02.01 : 2020.02.28 context by securityid

factorDoubleEMA 因子输出结果

同样基于前面的模拟原始数据，要计算4000只股票2020年2月份的分钟频 Alpha#1因子值，我们先用 select 语句生成如下数据表：

def alpha1SQL(t){
    res = select tradetime, securityid, mimax(pow(iif(ratios(close) - 1 < 0, mstd(ratios(close) - 1, 20), close), 2.0), 5) as val from t context by securityid
    return select tradetime, securityid, rank(val, percent=true) - 0.5 as val from res context by tradetime
}
input_alpha1 = select tradetime,securityid,close from loadTable("dfs://k_minute_level","k_minute") where date(tradetime) between 2020.02.01 : 2020.02.28
select top 10 * from input_alpha1
alpha1DDBSql = alpha1SQL(input_alpha1)

def alpha98SQL(mutable t){ //mutable：可变参数在函数体中可以被修改
  update t set adv5 = mavg(vol, 5), adv15 = mavg(vol, 15) context by securityid
  update t set rank_open = rank(X = open,percent=true), rank_adv15 =rank(X=adv15,percent=true) context by date(tradetime)
  update t set decay7 = mavg(mcorr(vwap, msum(adv5, 26), 5), 1..7), decay8 = mavg(mrank(9 - mimin(mcorr(rank_open, rank_adv15, 21), 9), true, 7), 1..8) context by securityid
  return select tradetime,securityid, `alpha98 as factorname, rank(X =decay7,percent=true)-rank(X =decay8,percent=true) as val from t context by date(tradetime)
}
input_alpha98 = select tradetime,securityid, vwap,vol,open from loadTable("dfs://k_minute_level","k_minute") where date(tradetime) between 2020.02.01 : 2020.02.28
alpha98DDBSql = alpha98SQL(input_alpha98)