https://code.kx.com/q/learn/ - click on 'download' - its a bad link

Hi,

I found some typos in docs/docs/style/new-wine.md.
line 78: amll -> small
line 92: lgical -> logical

Thanks for your work. Very informative.

Hi team, I need to pub sub using pyq from kdb . Can you please let me know an example of how it can be done . Can't find any example on web or otherwise

Parted is Arthur’s original terminology, and helps avoid confusion with partitioned database.

As per Behavior of asc table there appears to be undocumented functionality of asc on a table with a single key and single column:

But when there is only one non-keyed column, it's actually the sorted attribute that is set.

q)meta asc ([c1:`a`b] c2:2 1)
c | t f a
--| -----
c1| s 
c2| j   s
q)meta asc ([c1:`a`b] c2:2 1; c3:01b)
c | t f a
--| -----
c1| s
c2| j   p
c3| b

Documentation to be updated: https://code.kx.com/q/ref/asc/#asc
Duplicate information: https://code.kx.com/q/ref/set-attribute/#sorted

FYI: @StephenTaylor-Kx, CC: @rianoc

Issue

In several locations across code.kx.com we deal with finding all permutations. e.x.

• https://code.kx.com/q/learn/reading/strings/#its-more-fun-to-permute
• https://code.kx.com/phrases/math/#permutations

In both of these cases, I find that duplicate permutations are just completely unaddressed. If we are dealing with a set, naturally there will be no problem with duplicates. But the phrasebook does not clearly specify that we are dealing with sets, and the reading room puzzle goes further to specify our input is a string (clearly not a set). Furthermore, even if we assume that we are dealing with permutations of a set, we should ensure that the official solutions we provide handle these in a standardized and expected manner.

Clearly this is not the case.

Solutions

We should either:

1. Clarify that the solutions provided only deal with unique elements, and have undefined behaviour otherwise
2. Standardize the solutions to remove all duplicates
3. Standardize the solutions to enumerate all permutations, regardless of duplicates

We can also do any combination of these (e.x. solution (1) for phrasebook and (2) for the reading room problem).

My Contribution

I believe solution 2 is the best option. A list of permutations with duplicates seems redundant and unnecessary. You might as well distinct the output, in which case we should do this in the solutions provided. I don't know if applying distinct is necessarily the best solution here, in fact I have little experience with q, but it seems the most straightforward solution.

Phrasebook

• Before: {(1 0#x){raze(1 rotate)scan'x,'y}/x}
• After: {(1 0#x){distinct raze(1 rotate)scan'x,'y}/x}

Reading room cats cradle solution 1

This solution is actually incomplete and incorrect which is a separate issue of its own that caused this entire investigation. I have done the best I can to fix the solution and will provide it as a PR after discussion of this ticket.

• Before: {x {flip[y]where x=sum y}[s;] s vs til"j"$s xexp s:count x}, incorrect
• After: {distinct x @ b where all each (til c) in/:b:flip c vs til "j"$c xexp c:count x}

Reading room cats cradle solution 2

• Before: {(1 0#x) {raze({raze reverse 0 1 _ x}\)each x,'y}/ x}, strongly based off the phrasebook solution
• After: {(1 0#x) {distinct raze({raze reverse 0 1 _ x}\)each x,'y}/ x}

Ref: https://code.kx.com/q/wp/java-api/

This line

The API is contained in a single source file on GitHub.

is referencing code thats moved, so the link is broken.
Prob better if it just linked to the repo on github (https://github.com/KxSystems/javakdb) so that they can see details of the API/etc when visiting rather than bypassing it, and also allowing code to be refactored further.

Go to https://code.kx.com/q/learn/
On the big square links in centre of page, click the last one 'advanced q' ... it opens an empty page

https://code.kx.com/q/learn/blogs/ - "Kdb+ and Python" link broken

Go to https://code.kx.com/q/learn/ - click on 'Mountain tour'. The first link is is for 'TLDR'. Looks like a bad html instead of a title.

https://code.kx.com/q/interfaces/capiref/#sd0-remove-callback

The instructions for sd0 and sd0x say that they can return null under certain conditions.
Though the functions are stated as not returning anything on their signature (return void not void*)

https://code.kx.com/q/learn/ click on 'KX developer'. Takes you to https://code.kx.com/developer/. The menu along the top doesnt match the rest of the site.

• There appears to a parsing error with the new use of <pre><code class="language-q"> in the definition of code blocks over the previous use of triple back tick q as a code block start.

• In particular the definition of code containing ` as a character for the definition of a symbol list or in embedPy for accessing python objects is not being parsed in accordance with expected behaviour for q

The code blocks in this section show this incorrect parsing (this is the only instance observed so far)

https://code.kx.com/q/ml/automl/faq/#keras-models

Namely the definition of i.keraslist should be a list of symbols, and the definition of layer1_nm in customregkeras should also be a symbol list.

https://code.kx.com/q/basics/ipc/#compression
"The compression/decompression algorithms are proprietary and implemented as the z and u methods in c.java"

Should prob be
"The compression/decompression algorithms are proprietary and implemented as the compress and uncompress methods in c.java"

Background:
https://github.com/KxSystems/javakdb/blob/c9afe6fa32d7d3e3cddabdc9bd43f0155a5d2a1b/src/kx/c.java
Original has z/u methods

changed as part of 'reformatted, documented and provided simple examples' Jun 16 2017
KxSystems/javakdb@5e1ee8f#diff-6ad7d9db62e45794ad1ae4af83abecdd

Now uses 'compress' and 'uncompress' methods
https://github.com/KxSystems/javakdb/blob/master/src/kx/c.java

New location of C# API is https://github.com/KxSystems/csharpkdb

This doc https://code.kx.com/q/interfaces/csharp-client-for-q/ currently points to https://github.com/KxSystems/kdb/blob/master/c/c.cs in the first sentance. Should be changed to https://github.com/KxSystems/csharpkdb

Thanks