PyKDB provides two modules to call kdb+ from python and vice-versa designed for both efficiency and flexibility. In its core PyKDB is just a wrapper around k.h but it can automatically convert kdb+ objects to numpy/uuid/pandas objects. This makes it quite flexible to interact between kdb+ and python. Also pykdb is using boost.python and Boost.Python NumPy to enable seamless interoperability between kdb+ and the Python programming language. This also means that pykdb can be compiled on any platforms like win32/win64/macos and any unix systems.
$ python
Python 3.6.7 (default, Feb 28 2019, 07:28:18) [MSC v.1900 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import qip,numpy,pandas,uuid
>>> hdl = qip.open_connection(8888)
>>> hdl
996
>>> qip.execute(hdl,'1')
1
>>> qip.execute(hdl,'1f')
1.0
>>> qip.execute(hdl,'til 10')
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype=int64)
>>> qip.execute(hdl,'{til x}',10)
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype=int64)
>>> vec = qip.execute(hdl,'{til x}',10)
>>> vec = qip.execute(hdl,'{x + til count x}',vec)
>>> vec
array([ 0, 2, 4, 6, 8, 10, 12, 14, 16, 18], dtype=int64)
>>> qip.execute(hdl,'.z.D + til 10 ')
array(['2019-06-12', '2019-06-13', '2019-06-14', '2019-06-15',
'2019-06-16', '2019-06-17', '2019-06-18', '2019-06-19',
'2019-06-20', '2019-06-21'], dtype='datetime64[D]')
>>> qip.execute(hdl,'.z.P + til 10 ')
array(['2019-06-12T12:30:03.909487000', '2019-06-12T12:30:03.909487001',
'2019-06-12T12:30:03.909487002', '2019-06-12T12:30:03.909487003',
'2019-06-12T12:30:03.909487004', '2019-06-12T12:30:03.909487005',
'2019-06-12T12:30:03.909487006', '2019-06-12T12:30:03.909487007',
'2019-06-12T12:30:03.909487008', '2019-06-12T12:30:03.909487009'], dtype='datetime64[ns]')
>>> qip.execute(hdl,'.z.t + til 10 ')
array([37807949, 37807950, 37807951, 37807952, 37807953, 37807954,
37807955, 37807956, 37807957, 37807958], dtype='timedelta64[ms]')
>>> qip.execute(hdl,'`a`b! 1 2')
{'a': 1, 'b': 2}
>>> qip.execute(hdl,'`a`b! (til 10; 2)')
{'a': array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype=int64), 'b': 2}
>>> qip.execute(hdl,'flip `a`b! (til 10; 3+til 10)')
a b
0 0 3
1 1 4
2 2 5
3 3 6
4 4 7
5 5 8
6 6 9
7 7 10
8 8 11
9 9 12
>>> tbl = qip.execute(hdl,'flip `a`b! (til 10; 3+til 10)')
>>> qip.execute(hdl,'{x}',tbl)
a b
0 0 3
1 1 4
2 2 5
3 3 6
4 4 7
5 5 8
6 6 9
7 7 10
8 8 11
9 9 12
$ C:\q\w32\q.exe
KDB+ 3.6 2018.12.24 Copyright (C) 1993-2018 Kx Systems
w32/ 24()core 4095MB kuent desktop-u2c0ace 192.168.56.1 NONEXPIRE
q)piq:`$"C:\\q\\nlanguage\\w32\\piq"
q)py_initialize:piq 2:(`py_initialize;1);
q)set0:piq 2:(`set;2);
q)get0:piq 2:(`get;1);
q)eval3:piq 2:(`eval3;1);
q)exec3:piq 2:(`exec3;1);
q)py_initialize[];
q)exec3["import numpy,pandas,uuid"];
q)0N!enlist[;a;b] a ~ b:get0 set0["abc"] a:1000i;
(1b;1000i;1000i)
q)0N!enlist[;a;b] a ~ b:get0 set0["abc"] a:1000f;
(1b;1000f;1000f)
q)0N!enlist[;a;b] a ~ b:get0 set0["abc"] a:til 3;
(1b;0 1 2;0 1 2)
q)0N!enlist[;a;b] a ~ b:get0 set0["abc"] a:flip `a`bc!(1 2;2 2);
(1b;+`a`bc!(1 2;2 2);+`a`bc!(1 2;2 2))
Licensed under MIT License.