[unterrainer_ale]

Hi Jonathan, great, glad it worked. Just remember, you can send queries via parse trees like

handle(function;argument1;...;argumentN)

q)h(insert;`emulation;(1b;1b;1b))
,1
q)h "emulation"
a b c
-----
1 1 1
1 1 1

• This reply was modified 4 days, 1 hour ago by  unterrainer_ale.

[unterrainer_ale]

This is because the standard definition of .z.pg (the handle for synchronous messages) is value. value evaluates a string using variables that are defined locally. Because the values you use in the string query aren’t defined on the process you send the query to, you’re getting an error. You can send your query as a parse tree and it will use the local definition of the variables (from the client side) 

// server 
q)emulation:([] a:`boolean$(); b:`boolean$(); c:`boolean$())<br>q)a:1b;b:1b;c:1b;<br>q)a<br>1b<br>q)b<br>1b<br>q)c<br>1b<br>// client<br>q)value "2+2"<br>4<br>// You get the same error if you try to use value with variables that are not defined<br>q)value "g+h"<br>'g<br>  [1]  g+h<br>       ^<br>q)h "`emulation insert (1b;1b;1b)"
,0
q)h "emulation"
a b c
-----
1 1 1
q)h(insert;`emulation;(1b;1b;1b))<br>,1<br>q)h "emulation"<br>a b c<br>-----<br>1 1 1<br>1 1 1<br>// Note a,b,c is defined on the server<br>q)h"`emulation insert (a;b;c)"
,2
q)h"emulation"
a b c
-----
1 1 1
1 1 1
1 1 1
q)a:4
q)b:5
// a and b defined in local process (no boolean) 
q)value "a+b"
9
// defining variables locally that can be send via a parse tree
q)x:1b;y:1b;z:1b
q)h(insert;`emulation;(x;y;z))
,3
q)h"emulation"
a b c
-----
1 1 1
1 1 1
1 1 1
1 1 1

You can read more about IPC here https://code.kx.com/q/basics/ipc/#send-messages

code.kx.com

Interprocess communication | Developing | kdb+ and q documentation - kdb+ and q documentation

TCP/IP is used for communicating between processes. The protocol is extremely simple, as is the message format.

[unterrainer_ale]

If you look at your dictionary you can see that your key is a symbol atom, so you have to index by a symbol. Once you have done that, you can pass the value as parameter. This syntax should work

// Your dictionary
q).test.ex
  | ::
t1| {[x] x}
t2| {[x] x}
q).test.ex[`t1`t2]@'3 4
3 4

[unterrainer_ale]

You can create an untyped column by using below syntax:

t:([] a:(); b:(); c:()) 

However, the moment you get the first insert the types will be inferred from the first row you insert. Every successive insert that doesn’t align to the types of the first row will throw a type error. Apart from that, it’s pretty much one of the worst design decisions you can make. The power and speed comes from having typed columns and knowing how much memory needs to be assigned and knowing how big each memory block needs to be. Why not simply define the column as float? You can read about it here https://code.kx.com/q4m3/8_Tables/#82-empty-tables-and-schema

q)meta t
c| t f a
-| -----
a|
b|
c|
q)`t insert (1;2;4.0)</p><p><span style="font-size: 1rem;">,0</span><br></p><p>q)t</p><p><span style="font-size: 1rem;">a b c</span><br></p><p>-----</p><p>1 2 4</p><p>q)meta t</p><p><span style="font-size: 1rem;">c| t f a</span><br></p><p>-| -----</p><p>a| j</p><p>b| j</p><p>c| f</p><p>q)`t insert (1;2;3)
'type
  [0]  `t insert (1;2;3)
       ^

• This reply was modified 3 weeks, 2 days ago by  unterrainer_ale.

code.kx.com

Tables | Q For Mortals | A textbook forkdb+ and the q programming language - Q for Mortals

How to construct and handle kdb+ tables in the q programming language

[unterrainer_ale]

Reference for .Q.id

https://code.kx.com/q/ref/dotq/#id-sanitize

code.kx.com

The .Q namespace – tools | Reference | kdb+ and q documentation - kdb+ and q documentation

The .Q namespace contains utility objects for q programming

[unterrainer_ale]

The syntax for xcols is `newName xcol tableName , because it looks like you are sending the query over a handle you have to do either h_feed “`newName xcol tableName” or h_feed(xcol;`newName;tableName)

Note that xcol renames the first column to the first symbol. so if your update_date is at position n you need n-1 column names before on the left side. Not sure if that makes sense but you can read the documentation about xcol here https://code.kx.com/q/ref/cols/#xcol

You can also use .Q.id to sanitise your column names

(base) alexanderunterrainer@Alexanders-Laptop:~|⇒ cat table.csv

“”update_date*””,goodName

1,2

3,4

5,6

q)t:(“II”;enlist csv) 0:`table.csv

q)t

“update_date*” goodName

———————–

1 2

3 4

5 6

q).Q.id t

update_date goodName

——————–

1 2

3 4

5 6

or using xcol

q)`newGoodName xcol t

newGoodName goodName

——————–

1 2

3 4

5 6

code.kx.com

cols, xcol, xcols | Reference | kdb+ and q documentation - kdb+ and q documentation

cols, xcol and xcols are q keywords. cols returns the column names of a table. xcol renames tablecolumns. xcols reorders table columns.

[unterrainer_ale]

I would add some comments or descriptions for the parameter inputs to the functions for those who don’t have a finance background or are not familiar with the Kelly Criterion.

[unterrainer_ale]

As per the documentation, the par.txt file should contain the path to the individual segments. I don’t think you can use environment variables. https://code.kx.com/q4m3/14_Introduction_to_Kdb%2B/#144-segmented-tables

“<code style=”display: inline !important;”>par.txt

containing the paths of the physical locations of the segments, one segment path per line.”

code.kx.com

14. Introduction to Kdb+ - Q for Mortals

Q for Mortals: a textbook for kdb+ and the q programming language

[unterrainer_ale]

Hi fav_pa,

yes, a window join wj for several columns is possible. The syntax is

wj[w;c;t;(q;(f0;c0);(f1;c1))]

where w is your window, c are the columns you want to match on, in your case this would be sym, date time and trader type. As with the asof join, aj, the time column should be the last in your list, so I assume it’s sym,traderType,datetime (I am avoiding the symbol syntax because of the known UI issue).

You can read more about wj here https://code.kx.com/q4m3/9_Queries_q-sql/#999-window-join

code.kx.com

9. Queries – q-sql - Q for Mortals

Q for Mortals: a textbook for kdb+ and the q programming language

[unterrainer_ale]

@laura would you mind formatting my answer? thanks

[unterrainer_ale]

Hi rgiu70,

not sure if this is what you are looking for, but if you have a table and just want to get the open, high, low, close by sym you can run the following query

</p><p>q)t:([] sym:20?AAPLMSFTC; price:20?100.0)

q)t

sym price

————-

MSFT 39.27524

MSFT 51.70911

C 51.59796

AAPL 40.66642

MSFT 17.80839

C 30.17723

MSFT 78.5033

C 53.47096

MSFT 71.11716

AAPL 41.1597

AAPL 49.31835

MSFT 57.85203

C 8.388858

MSFT 19.59907

C 37.5638

AAPL 61.37452

AAPL 52.94808

C 69.16099

AAPL 22.96615

MSFT 69.19531

q)exec ohlc!(first;max;min;last)@\:price by sym from t

| o h l c

—-| ———————————–

AAPL| 40.66642 61.37452 22.96615 22.96615

C | 51.59796 69.16099 8.388858 69.16099

MSFT| 39.27524 78.5033 17.80839 69.19531

q)

`

[unterrainer_ale]

Hi eohara_kdb, have you tried the same with a table in a q process and does this error only occur on a HDB?

[unterrainer_ale]

Hi,

I believe this is what you are looking for.

q)t:([] sym:`a`a`a`b`b`c`c`a`b`c;price:1 1 2 3 4 5 5 8 9 7)

q)t

sym price

---------
a 1
a 1
a 2
b 3
b 4
c 5
c 5
a 8
b 9
c 7

q)select from t where (differ;price) fby sym

sym price

---------
a 1
a 2
b 3
b 4
c 5
a 8
b 9
c 7

 

You can read more about fby here fby – Reference – kdb+ and q documentation – kdb+ and q documentation (kx.com)

code.kx.com

kdb+ and q documentation

Documentation for kdb+ and the q programming language

[unterrainer_ale]

Hi,

I just use terminal on my Mac to start a q process and then either directly work in the console or use a IDE QStudio, QPad, kdbstudio, IntelliJ or VS Code to connect to it. That works fine. You can do the same on Windows using CMD and the IDEs

[unterrainer_ale]

I just replicated the error myself, so it looks like indeed, when you use the same name for the parameter of your functional select as the virtual column (date in this case) it doesn’t work. What you should keep in mind is that when you are on a hdb, the virtual column, date is actually loaded into memory, meaning, the variable date exists. If you just run date on your HDB it will output all available dates. This also impacts your second question, where you asked about count 0. You have an enlist in your where clause (enlist date), and on the HDB I tested, it actually returned 1 single record. If I run the same with enlist dt, it throws a length error