[mauricelim]

What you need is actually the (?) vector conditional operator: https://code.kx.com/q/ref/vector-conditional/

? accepts a boolean list as the first argument but $ accepts a boolean atom. Think of ? as a boolean mask.

[mauricelim]

You can use the drop (_) operator, where n is a number to drop.

n_list

To remove from the end of the list:

-n_list neg[n]_list

If you want to remove nulls, you can do something like:

list except 0N

You can use the drop with each (‘) operator to remove the same numbers of items from each list of different length, i.e.

n_'(ls;ls2) -n_'(ls;ls2) neg[n]_'(ls;ls2)

Or removing nulls from them

(ls;ls2)except'0N

 

[mauricelim]

The range of generate for symbols is only from abcdefghijklmnop as documented here: https://code.kx.com/q/ref/deal/#roll-and-deal

As to why is that so, I have no idea as well.

But you can use something like this to solve your problem:

q){[num;len] neg[num]?`${.Q.a cross x}/[len-1;enlist each .Q.a]}[17;1] `e`n`d`g`r`y`b`p`j`l`s`i`f`k`w`u`c 
q)asc{[num;len] neg[num]?`${.Q.a cross x}/[len-1;enlist each .Q.a]}[256;2] `s#`af`an`aq`as`ax`ay`ba`bd`be`bj`bk`bl`bn`bo`bu`bw`ca`cb`cc`cd`cf`cn`co`cq`c..

 

[mauricelim]

You could do something like this:

q)fix:{.Q.fmt'[x+1+count each string floor y;x;y]} 
q)select time,sym,fix[1]price from trade 
time sym price 
------------------------ 
09:30:00.000 a "10.8" 
09:31:00.000 a "11.8" 
09:32:00.000 a "13.2" 
09:30:00.000 b "100.8" 
09:31:00.000 b "107.0" 
09:32:00.000 b "124.0" 

q)select time,sym,"F"$fix[1]price from trade 
time       sym price 
---------------------- 
09:30:00.000 a 10.8 
09:31:00.000 a 11.8 
09:32:00.000 a 13.2 
09:30:00.000 b 100.8 
09:31:00.000 b 107 
09:32:00.000 b 124

Reference: https://code.kx.com/q/ref/dotq/#qfmt-format

Note that if you do cast it back to a float, the decimal does not show for .0

[mauricelim]

Yes, it is creating a table schema of specific type for the cols. So when you are manipulating the table, it throws a type error if the type is not of that defined in the schema.

https://code.kx.com/q/ref/insert/#type

More information here: https://code.kx.com/q4m3/8_Tables/#82-empty-tables-and-schema

[mauricelim]

1. : represents each left and $ is casting.

https://code.kx.com/q/ref/maps/#each-left-and-each-right

https://code.kx.com/q4m3/7_Transforming_Data/#72-cast

So for the line “nsfi”$:() it is actually casting an empty list with every type on the left (hence each left) so that it creates a list of that specific type. You can use .Q.s1 to view the string representation of the object.

https://code.kx.com/q/ref/dotq/#qs1-string-representation

q)"n"$() 
`timespan$() 
q).Q.s1 "nsfi"$:() 
"(`timespan$();`symbol$();`float$();`int$())"

2. bang (!) is used to create a dictionay with the keys on the left and values on the right with the correct length of list. Then flipping a dict will turn it into a tbl.

https://code.kx.com/q/ref/dict/

https://code.kx.com/q/kb/faq/#flip-a-column-dictionary

3. 2! enkeys the table by the first 2 cols.

https://code.kx.com/q/ref/enkey/

Hope this helps!