[fbodon]

You can sort the table column by column using iasc. This approach is described here – see function disksort.

https://code.kx.com/q/wp/intraday-writedown/#wq

[fbodon]

You might find this document useful: https://dataintellect.com/blog/adventure-in-retrieving-memory-size-of-kdb-object/ It covers the memory need of attributes.

[fbodon]

This is implicit composition (‘). You can see this from the parse tree:

 

q)parse "(3#,:)" 
' 
(#;3) 
,:

 

In k, the expression ,: corresponds to enlist. This construct is the same as repetition (*) in Python.

See: https://code.kx.com/q/ref/compose/

[fbodon]

This 1 f pattern is crazy, but i love it

[fbodon]

I would do

@[-6_string .z.p; 4 7 10; :; “–T”]

or if you prefer narrow square brackets then

@[; 4 7 10; :; “–T”] -6_string .z.p

[fbodon]

Haha. Nice! The good old datetime! I already erased it from my memory

[fbodon]

To avoid user queries blocking each other, you can set up a pool of q processes (HDBs) and use socket sharding as per https://code.kx.com/q/wp/socket-sharding/ or a TCP load balancer.