QUANTAXIS的数据结构

datastruct

属性用@property装饰器装饰,进行懒运算 提高效率

QA_DataStruct具有的功能:

  • 数据容器

  • 数据变换 [分拆/合并/倒序] split/merge

  • 数据透视 pivot

  • 数据筛选 selects/select_time/select_time_with_gap/select_code/get_bar/select_month

  • 数据复权 to_qfq/to_hfq

  • 数据显示 show

  • 格式变换 to_json/to_pandas/to_list/to_numpy/to_hdf

  • 数据库式查询 query

  • 画图 plot

  • 计算指标 add_func

  • 生成器 panel_gen(按时间分类的面板生成器)/security_gen(按股票分类的股票生成器)

QA_DataStruct_Stock_block

  • (属性)该类下的所有板块名称 block_name

  • 查询某一只股票所在的所有板块 get_code(code)

  • 查询某一个/多个板块下的所有股票 get_block(blockname)

  • 展示当前类下的所有数据 show

我们可以通过

import QUANTAXIS as QA
# QA.QA_fetch_stock_day_adv
# QA.QA_fetch_stock_min_adv
# QA.QA_fetch_index_day_adv
# QA.QA_fetch_index_min_adv

day线的参数是code, start, end min线的参数是code, start, end, frequence='1min'

其中 code 可以是一个股票,也可以是一列股票(list)

1. 取一个股票的数据

QA.QA_fetch_stock_day_adv('000001','2017-01-01','2017-10-01')
In [5]: QA.QA_fetch_stock_day_adv('000001','2017-01-01','2017-10-01')
Out[5]: QA_DataStruct_Stock_day with 1 securities

2. 取多个股票的数据

QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-01-01','2017-10-01')
In [6]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-01-01','2017-10-01')
Out[6]: QA_DataStruct_Stock_day with 2 securities

3. 显示结构体的数据 .data

In [10]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').data
Out[10]:
code open high low close volume date
date code
2017-09-20 000001 000001 11.14 11.37 11.05 11.29 787154.0 2017-09-20
2017-09-21 000001 000001 11.26 11.51 11.20 11.46 692407.0 2017-09-21
2017-09-22 000001 000001 11.43 11.52 11.31 11.44 593927.0 2017-09-22
2017-09-25 000001 000001 11.44 11.45 11.18 11.29 532391.0 2017-09-25
2017-09-26 000001 000001 11.26 11.30 10.96 11.05 967460.0 2017-09-26
2017-09-27 000001 000001 11.01 11.08 10.90 10.93 727188.0 2017-09-27
2017-09-28 000001 000001 10.98 10.98 10.82 10.88 517220.0 2017-09-28
2017-09-29 000001 000001 10.92 11.16 10.86 11.11 682280.0 2017-09-29
2017-09-20 000002 000002 28.50 29.55 28.00 28.73 613095.0 2017-09-20
2017-09-21 000002 000002 28.50 29.06 27.75 28.40 536324.0 2017-09-21
2017-09-22 000002 000002 28.39 28.67 27.52 27.81 423093.0 2017-09-22
2017-09-25 000002 000002 27.20 27.20 26.10 26.12 722702.0 2017-09-25
2017-09-26 000002 000002 26.12 27.22 26.10 26.76 593044.0 2017-09-26
2017-09-27 000002 000002 27.00 27.28 26.52 26.84 367534.0 2017-09-27
2017-09-28 000002 000002 27.00 27.15 26.40 26.41 262347.0 2017-09-28
2017-09-29 000002 000002 26.56 26.80 26.00 26.25 345752.0 2017-09-29

4. 显示结构体的开/高/收/低 .open/.high/.close/.low

In [5]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').high
Out[5]:
date code
2017-09-20 000001 11.37
2017-09-21 000001 11.51
2017-09-22 000001 11.52
2017-09-25 000001 11.45
2017-09-26 000001 11.30
2017-09-27 000001 11.08
2017-09-28 000001 10.98
2017-09-29 000001 11.16
2017-09-20 000002 29.55
2017-09-21 000002 29.06
2017-09-22 000002 28.67
2017-09-25 000002 27.20
2017-09-26 000002 27.22
2017-09-27 000002 27.28
2017-09-28 000002 27.15
2017-09-29 000002 26.80
Name: high, dtype: float64

5. 结构体拆分 splits()

当一个DataStruct里面存在多个证券时,可以通过拆分的方法,将其变成多个DataStruct

In [3]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').splits()
Out[3]:
[< QA_DataStruct_Stock_day with 1 securities >,
< QA_DataStruct_Stock_day with 1 securities >]

6. 数据结构复权to_qfq()/to_hfq()

返回的是一个DataStruct,用.data展示返回的数据的结构

其中DataStruct.if_fq的属性会改变

In [4]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').to_qfq().data
Out[4]:
code open high low close volume date \
date code
2017-09-20 000001 000001 11.14 11.37 11.05 11.29 787154.0 2017-09-20
2017-09-21 000001 000001 11.26 11.51 11.20 11.46 692407.0 2017-09-21
2017-09-22 000001 000001 11.43 11.52 11.31 11.44 593927.0 2017-09-22
2017-09-25 000001 000001 11.44 11.45 11.18 11.29 532391.0 2017-09-25
2017-09-26 000001 000001 11.26 11.30 10.96 11.05 967460.0 2017-09-26
2017-09-27 000001 000001 11.01 11.08 10.90 10.93 727188.0 2017-09-27
2017-09-28 000001 000001 10.98 10.98 10.82 10.88 517220.0 2017-09-28
2017-09-29 000001 000001 10.92 11.16 10.86 11.11 682280.0 2017-09-29
2017-09-20 000002 000002 28.50 29.55 28.00 28.73 613095.0 2017-09-20
2017-09-21 000002 000002 28.50 29.06 27.75 28.40 536324.0 2017-09-21
2017-09-22 000002 000002 28.39 28.67 27.52 27.81 423093.0 2017-09-22
2017-09-25 000002 000002 27.20 27.20 26.10 26.12 722702.0 2017-09-25
2017-09-26 000002 000002 26.12 27.22 26.10 26.76 593044.0 2017-09-26
2017-09-27 000002 000002 27.00 27.28 26.52 26.84 367534.0 2017-09-27
2017-09-28 000002 000002 27.00 27.15 26.40 26.41 262347.0 2017-09-28
2017-09-29 000002 000002 26.56 26.80 26.00 26.25 345752.0 2017-09-29
preclose adj
date code
2017-09-20 000001 NaN 1.0
2017-09-21 000001 11.29 1.0
2017-09-22 000001 11.46 1.0
2017-09-25 000001 11.44 1.0
2017-09-26 000001 11.29 1.0
2017-09-27 000001 11.05 1.0
2017-09-28 000001 10.93 1.0
2017-09-29 000001 10.88 1.0
2017-09-20 000002 NaN 1.0
2017-09-21 000002 28.73 1.0
2017-09-22 000002 28.40 1.0
2017-09-25 000002 27.81 1.0
2017-09-26 000002 26.12 1.0
2017-09-27 000002 26.76 1.0
2017-09-28 000002 26.84 1.0
2017-09-29 000002 26.41 1.0

7. 数据透视 .pivot()

In [6]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').pivot('open')
Out[6]:
code 000001 000002
date
2017-09-20 11.14 28.50
2017-09-21 11.26 28.50
2017-09-22 11.43 28.39
2017-09-25 11.44 27.20
2017-09-26 11.26 26.12
2017-09-27 11.01 27.00
2017-09-28 10.98 27.00
2017-09-29 10.92 26.56

8. 数据的自定义筛选.selects(code,start,end)

In [4]: data.selects('000005','2018-06-01','2018-06-05').data
Out[4]:
open high low close volume amount
date code
2018-06-01 000005 3.42 3.46 3.39 3.44 47210.0 16151548.0
2018-06-04 000005 3.43 3.45 3.39 3.40 35039.0 11945360.0
2018-06-05 000005 3.39 3.42 3.38 3.42 34129.0 11609377.0

9. 数据的时间筛选.select_time(start,end)

In [10]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').select_time('2017-09-20','2017-09-25')
Out[10]: QA_DataStruct_Stock_day with 2 securities
In [11]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').select_time('2017-09-20','2017-09-25').data
Out[11]:
code open high low close volume date
date code
2017-09-20 000001 000001 11.14 11.37 11.05 11.29 787154.0 2017-09-20
2017-09-21 000001 000001 11.26 11.51 11.20 11.46 692407.0 2017-09-21
2017-09-22 000001 000001 11.43 11.52 11.31 11.44 593927.0 2017-09-22
2017-09-25 000001 000001 11.44 11.45 11.18 11.29 532391.0 2017-09-25
2017-09-20 000002 000002 28.50 29.55 28.00 28.73 613095.0 2017-09-20
2017-09-21 000002 000002 28.50 29.06 27.75 28.40 536324.0 2017-09-21
2017-09-22 000002 000002 28.39 28.67 27.52 27.81 423093.0 2017-09-22
2017-09-25 000002 000002 27.20 27.20 26.10 26.12 722702.0 2017-09-25

10. 数据按时间往前/往后推 select_time_with_gap(time,gap,methods)

time是你选择的时间 gap是长度 (int) methods有 '<=','lte','<','lt','eq','==','>','gt','>=','gte'的选项

In [14]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').select_time_with_gap('2017-09-20',2,'gt')
Out[14]: QA_DataStruct_Stock_day with 2 securities
In [15]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').select_time_with_gap('2017-09-20',2,'gt').data
Out[15]:
code open high low close volume date
date code
2017-09-21 000001 000001 11.26 11.51 11.20 11.46 692407.0 2017-09-21
2017-09-22 000001 000001 11.43 11.52 11.31 11.44 593927.0 2017-09-22
2017-09-21 000002 000002 28.50 29.06 27.75 28.40 536324.0 2017-09-21
2017-09-22 000002 000002 28.39 28.67 27.52 27.81 423093.0 2017-09-22

11. 选取某一个月份的数据 select_month(month)

可以通过select_month 来选取某一个月份的数据

In [4]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').select_month('2017-09')
Out[4]: < QA_DataStruct_Stock_day with 2 securities >

12. 选取结构组里面某一只股票select_code(code)

In [16]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').select_code('000001')
Out[16]: QA_DataStruct_Stock_day with 1 securities
In [17]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').select_code('000001').data
Out[17]:
code open high low close volume date
date code
2017-09-20 000001 000001 11.14 11.37 11.05 11.29 787154.0 2017-09-20
2017-09-21 000001 000001 11.26 11.51 11.20 11.46 692407.0 2017-09-21
2017-09-22 000001 000001 11.43 11.52 11.31 11.44 593927.0 2017-09-22
2017-09-25 000001 000001 11.44 11.45 11.18 11.29 532391.0 2017-09-25
2017-09-26 000001 000001 11.26 11.30 10.96 11.05 967460.0 2017-09-26
2017-09-27 000001 000001 11.01 11.08 10.90 10.93 727188.0 2017-09-27
2017-09-28 000001 000001 10.98 10.98 10.82 10.88 517220.0 2017-09-28
2017-09-29 000001 000001 10.92 11.16 10.86 11.11 682280.0 2017-09-29

13. 取某一只股票的某一个时间的bar(code,time,if_trade)

In [19]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').get_bar('000001','2017-09-20')
Out[19]:
code open high low close volume date
date code
2017-09-20 000001 000001 11.14 11.37 11.05 11.29 787154.0 2017-09-20

14. 统计学部分

14.1. 平均价 price

为了统计学指标的需要, price=AVERAGE(open+high+low+close)

price是一个pd.Series

In [7]: QA.QA_fetch_stock_day_adv('000001','2017-09-20','2017-10-01').price
Out[7]:
date code
2017-09-20 000001 11.2125
2017-09-21 000001 11.3575
2017-09-22 000001 11.4250
2017-09-25 000001 11.3400
2017-09-26 000001 11.1425
2017-09-27 000001 10.9800
2017-09-28 000001 10.9150
2017-09-29 000001 11.0125
dtype: float64

14.2. price均值 mean

mean是price的均值

In [6]: QA.QA_fetch_stock_day_adv('000001','2017-09-20','2017-10-01').mean
Out[6]: 11.173125

14.3. max/min

max/min 分别是price序列的最大值和最小值

In [8]: QA.QA_fetch_stock_day_adv('000001','2017-09-20','2017-10-01').max
Out[8]: 11.424999999999999
In [9]: QA.QA_fetch_stock_day_adv('000001','2017-09-20','2017-10-01').min
Out[9]: 10.915000000000001

14.4. 方差/样本方差 pvariance/variance

分别是price的方差和样本方差

In [10]: QA.QA_fetch_stock_day_adv('000001','2017-09-20','2017-10-01').variance
Out[10]: 0.0367852678571427
In [11]: QA.QA_fetch_stock_day_adv('000001','2017-09-20','2017-10-01').pvariance
Out[11]: 0.03218710937499986

14.5. 标准差/样本标准差 pstdev/stdev

分别是price的总体标准差和样本标准差

In [12]: QA.QA_fetch_stock_day_adv('000001','2017-09-20','2017-10-01').pstdev
Out[12]: 0.17940766253145338
In [13]: QA.QA_fetch_stock_day_adv('000001','2017-09-20','2017-10-01').stdev
Out[13]: 0.19179485878704544

14.6. 调和平均数 mean_harmonic

price的调和平均数

In [14]: QA.QA_fetch_stock_day_adv('000001','2017-09-20','2017-10-01').mean_harmonic
Out[14]: 11.170242242781745

14.7. 众数 mode

返回price的众数 (注意: price序列可能没有众数,因此可能会报错,内部处理后,返回None)

In [31]: QA.QA_fetch_stock_day_adv('000001','2017-01-20','2017-10-01').mode
Out[31]: 9.1375
In [31]: QA.QA_fetch_stock_day_adv('000001','2017-01-20','2017-10-01').mode
Out[31]: None

14.8. 振幅 amplitude

返回price的振幅

In [33]: QA.QA_fetch_stock_day_adv('000001','2017-01-20','2017-10-01').amplitude
Out[33]: 3.1325000000000003

14.9. 偏度 skew

返回price的偏度

In [35]: QA.QA_fetch_stock_day_adv('000001','2017-01-20','2017-10-01').skew
Out[35]: 0.70288041557825753

14.10. 峰度 kurt

返回price的峰度

In [37]: QA.QA_fetch_stock_day_adv('000001','2017-01-20','2017-10-01').kurt
Out[37]: -1.0703273213086726

14.11. 百分比变化 pct_change

返回price的百分比变化

In [40]: QA.QA_fetch_stock_day_adv('000001','2017-09-20','2017-10-01').pct_change
Out[40]:
date code
2017-09-20 000001 NaN
2017-09-21 000001 0.012932
2017-09-22 000001 0.005943
2017-09-25 000001 -0.007440
2017-09-26 000001 -0.017416
2017-09-27 000001 -0.014584
2017-09-28 000001 -0.005920
2017-09-29 000001 0.008933
dtype: float64

14.12. 平均绝对偏差 mad

返回price的平均绝对偏差

In [41]: QA.QA_fetch_stock_day_adv('000001','2017-09-20','2017-10-01').mad
Out[41]: 0.16062499999999957

14.13. 价格差分 price_diff

返回价格的一阶差分

In [42]: QA.QA_fetch_stock_day_adv('000001','2017-09-20','2017-10-01').price_diff
Out[42]:
date code
2017-09-20 000001 NaN
2017-09-21 000001 0.1450
2017-09-22 000001 0.0675
2017-09-25 000001 -0.0850
2017-09-26 000001 -0.1975
2017-09-27 000001 -0.1625
2017-09-28 000001 -0.0650
2017-09-29 000001 0.0975
dtype: float64

15. 画图 plot(code)

如果是()空值 就会把全部的股票都画出来

In [20]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').plot()
QUANTAXIS>> The Pic has been saved to your path: .\QA_stock_day_codepackage_bfq.html
In [21]: QA.QA_fetch_stock_day_adv(['000001','000002'],'2017-09-20','2017-10-01').plot('000001')
QUANTAXIS>> The Pic has been saved to your path: .\QA_stock_day_000001_bfq.html