Simulate a Default Argument With a Null Check

Question

Write a function topN that:

Takes a table t and an optional count n
Returns the first n rows of t, sorted by price descending
Defaults to 5 rows when n is not supplied (i.e. the caller passes 0N)

Solution

topN:{[t; n]
  n: $[null n; 5; n];
  select[n; >price] from t
 }

trade:([] sym:`AAPL`MSFT`GOOG`AAPL`MSFT; price:185.5 410.1 142.8 187.0 411.2; qty:100 200 300 150 250)

topN[trade; 3]    / top 3 by price
topN[trade; 0N]   / top 5 (default)

Explanation

$[null n; 5; n] is a three-part conditional: if n is null, return 5; otherwise return n as-is. This line always runs first so every subsequent line in the function can treat n as a guaranteed non-null value.

null returns 1b for any typed null: 0N (long null), 0n (float null), ` (symbol null), 0Np (timestamp null), and so on. This makes the pattern work regardless of what type the caller passes for the missing argument.

Alternative using fill (^):

topN:{[t; n]
  n: 5 ^ n;
  select[n; >price] from t
 }

5 ^ n replaces n with 5 if n is null. This is slightly shorter but less explicit — prefer $[null n; ...] when the logic is part of a longer function and you want it to read clearly.

Calling convention: By convention, callers signal "use the default" by passing the appropriate typed null. For a long argument, that is 0N; for a float, 0n; for a symbol, `. Document this in the function's header comment so callers know what to pass.