q cheat sheet

The operators the q problems lean on. Run them in a q session to see the exact output.

Task	q	Note
Running total	`sums x`	sums 1 2 3 gives 1 3 6
Consecutive differences	`deltas x`	the first item is kept as is
Moving average	`3 mavg x`	early items use what is available
Round down to a bucket	`5 xbar x`	5 xbar 7 12 gives 5 10
Unique values	`distinct x`	keeps first-seen order
Sort	`asc x / desc x`
Count per value	`count each group x`	returns a dict
Weighted average	`w wavg x`	weights go on the left
Group and aggregate	`select avg price by sym from t`	result is keyed by sym
Filter by group aggregate	`select from t where size=(max;size) fby sym`	rows with the max size per sym
As-of join	aj[`sym`time; trade; quote]	the last key column is the time
Window join	wj[w; `sym`time; t; (q; (max;`bid))]	aggregate over a time window
Apply over a list	`f each x`
Fold and scan	`+/ x and +\ x`	over folds, scan keeps each step