formatting a table with @ (apply) to use lists as a format type (for example symbol "s" is "S")

[simon_watson_sj]

Hi Team,

I'm wondering how to use the @ (apply) paradigm to  set a table column so that it accepts lists rather than atoms.

The below illustrates my problem. I can use @ (apply) to cast a column to an atomic data type but can't seem to manage to get it to a listed datatype.

// define an empy table
tbl:flip `ref`stamp`note`usr!(`symbol$();`timestamp$();"";())

//use apply to make column usr a symbol type:
@[tbl;`usr;{y$x};"s"]
// works (test using meta tbl)

//use apply to make column note a string type (nested character):
@[tbl;`usr;{y$x};"C"]
// does not work - result is a column of type c (single character) (test using meta tbl)

My question is:

Is there a way to use apply (@) to set a column in a table to a nested data type (so "C" for a string or "S" for a list of symbols for each record)?

Regards,

Simon

[rocuinneagain]

meta on an empty table will never show C. This is because kdb+ does not have a dedicated datatype for lists of lists. The empty list will be () which is of type 0h. On a populated table meta inspects the first item in a list of lists and uses it's type to populate it's result. It cannot do this on an empty table.

q)tb:([] a:1 2;b:("wo";"rd"))
q)tb
a b
------
1 "wo"
2 "rd"
q)meta tb
c| t f a
-| -----
a| j
b| C
q)delete from `tb
`tb
q)tb
a b
---
q)meta tb
c| t f a
-| -----
a| j
b|
q).Q.s1 tb
"+`a`b!(`long$();())"

[simon_watson_sj]

Thanks rocuinneagain - that's good insight.

Given the above constraint, I have the following issue:

// define a table with the intention of loading a string to a columnn. 
// as discussed, it is empty so 'uptar' is defined as having type "c". 
test: ([]test: `$(); uptar:"C"$());

// With this done, I can't see a valid upsert to get a string into the uptar column. 
test upsert ([]test:  `rta; uptar: enlist "ra");
test upsert ([]test: enlist `rta; uptar:  "ra");
test upsert ([]test: enlist `rta; uptar:enlist  enlist "ra");

// finally, this one projects - so works but still not correctly.
test upsert ([]test:  `rta; uptar:  "ra");

Is it possible to insert a nested list into an empty table or must we use a 'set' workaround?

Thanks and regards,

Simon

[davidcrossey · Accepted solution]

Hi Simon,

You can leave the type of the (future) string column as empty in the table definition and let q infer the type once the first upsert occurs:

q)test: ([]test: `$(); uptar:());
q)test upsert ([]test:  `rta; uptar: enlist "ra")
test uptar
----------
rta  "ra"
q)meta test upsert ([]test:  `rta; uptar: enlist "ra")
c    | t f a
-----| -----
test | s
uptar| C

Happy Christmas!

[simon_watson_sj]

I AM NOT WORTHY!!!! 😉

Great to hear from you David!

Thanks to you and rocuinneagain.

I appreciate both of you taking the time.

Both of you have a eGuinness on me.

🍻  (eDrinks: like real ones but rubbish)

Simon