PyQ - 将数据类型转换为空表

Question

在PyQ中，我可以创建一个字典并通过翻转来转置它以获得一个表...

q.set(':alpha', q('!', ["Name", "Ask", "Bid", "Time"], ()).flip)

但是，正如预期的那样，当我在q中加载它并检索元数据时，它将所有列的char类型显示为符号...

>>> q.get(':alpha').show()
Name  Ask              Bid            Time
---------------------------------------------------------------
"BTC" "16351.0"        "16350.0"      "2017-12-12T17:32:33.09"
"ETH" "589.89999993"   "585.25"       "2017-12-12T17:32:32.697"
"LTC" "297.0"          "296.29570358" "2017-12-12T17:32:32.353"
"BTC" "16355.0"        "16351.0"      "2017-12-12T17:32:44.777"
"ETH" "589.89999993"   "585.25"       "2017-12-12T17:32:42.15"
"LTC" "297.99940398"   "296.29570359" "2017-12-12T17:32:44.433"
"BTC" "16359.99999998" "16350.0"      "2017-12-12T17:32:53.713"
"ETH" "589.89999993"   "585.2500001"  "2017-12-12T17:32:53.197"
"LTC" "297.99940398"   "295.0"        "2017-12-12T17:32:51.37"
"BTC" "16355.0"        "16350.0"      "2017-12-12T17:33:02.433"
"ETH" "585.2500001"    "585.25"       "2017-12-12T17:33:03.497"
"LTC" "297.99940397"   "295.0"        "2017-12-12T17:33:01.463"
>>> q()
q)\l alpha
`alpha
q)meta alpha
c   | t f a
----| -----
Name| s
Ask | s
Bid | s
Time| s
q)

我相信当我执行以下查询时，控制台会引发'type错误的原因...

q)select Name, max Ask, max Bid, Time from alpha
'type
q)select max Ask, max Bid, Time by Name from alpha
'type
q)select from alpha where Bid=(max;Bid) fby Name
'type
q)select from alpha where Name=`BTC
Name Ask Bid Time
-----------------

我知道在q中我可以发出以下命令来实现这一点......

q)alpha:([]Name:`symbol$(); Ask:`float$(); Bid:`float$(); Time:`datetime$())
q)meta alpha
c   | t f a
----| -----
Name| s
Ask | f
Bid | f
Time| z
q)

为PyQ中的每一列定义char类型的正确语法是什么？

谢谢！

Answer 1

您也可以在PyQ中指定类型：

>>> from pyq import K
>>> q('!', ["Name", "Ask", "Bid", "Time"], [K.symbol([]), K.float([]), K.float([]), K.timestampt([])]).flip.meta.show()
c   | t f a
----| -----
Name| s
Ask | f
Bid | f
Time| p

或者在您的情况下，您可以使用以下内容：

>>> q('!', ["Name", "Ask", "Bid", "Time"], [K.symbol(["LTC-USD","BTC-USD"]), 
... K.float([310.,16700.]), 
... K.float([310.01,16700.92]), 
... K.timestamp([datetime.datetime(2017,12,13,17,40,44), datetime.datetime(2017,12,13,17,40,45)])]).flip.show()
Name    Ask   Bid      Time
----------------------------------------------------
LTC-USD 310   310.01   2017.12.13D17:40:44.000000000
BTC-USD 16700 16700.92 2017.12.13D17:40:45.000000000

上面的例子可以进一步简化，因为PyQ知道如何将Python类型转换为q类型。

>>> q('!', ["Name", "Ask", "Bid", "Time"], [["LTC-USD","BTC-USD"], [310.,16700.], [310.01,16700.92], 
... [datetime.datetime(2017,12,13,17,40,44), datetime.datetime(2017,12,13,17,40,45)]]).flip.meta.show()
c   | t f a
----| -----
Name| s
Ask | f
Bid | f
Time| p

您可以在PyQ用户指南的Constructs and casts部分阅读更多内容。

Answer 2

我将datetime char类型传递给timestamp列。在q中，我将这些定义为......

q)table:([]date:`datetime$(); name:`symbol$())
q)meta table
c   | t f a
----| -----
date| z
name| s

q)table:([]date:`timestamp$(); name:`symbol$())
q)meta table
c   | t f a
----| -----
date| p
name| s

通过将日期时间解析为时间戳并转换为K.timestamp([])，我能够解决这个问题...

代码如下......

from bittrex.bittrex import Bittrex, API_V2_0
from datetime import datetime
import time
from pyq import q, K

get_bittrex = Bittrex(None, None)
starttime = time.time()

q.load(':alpha')

while True:
    market_result = get_bittrex.get_market_summaries()['result']
    for res in market_result:
        market_name = res['MarketName']
        ask = float(res['Ask'])
        bid = float(res['Bid'])
        last = float(res['Last'])
        volume = float(res['Volume'])
        dt = res['TimeStamp']
        if market_name in ['USDT-BTC', 'USDT-ETH', 'USDT-LTC', 'USDT-XRP', 'USDT-NEO', 'USDT-BCC', 'USDT-ZEC', 'USDT-XMR', 'USDT-DASH']:
            ts = datetime.strptime(dt, '%Y-%m-%dT%H:%M:%S.%f')
            data = [market_name[5:], ask, bid, last, volume, ts]
            q.upsert(':alpha', [data])
            q.get(':alpha').show()
    time.sleep(60.0)

现在返回我表中正确的char类型...

q)meta alpha
c   | t f a
----| -----
Name| s
Ask | f
Bid | f
Last| f
Vol | f
Time| p