Longest Low-Slippage Execution Window

Each execution log row is [timestamp, venue, fillPrice, benchmarkPrice, volume], with integer numeric fields encoded as strings. Filter to targetVenue, then sort by increasing timestamp; equal timestamps retain input order.
The slippage cost of a row is abs(fillPrice - benchmarkPrice) * volume. Return the maximum length of a contiguous filtered window whose fill-price range is at most maxPriceRange and whose total slippage cost is at most maxSlippage. Return 0 if no target row is valid.

Function
longestExecutionWindow(logs: String[][], targetVenue: String, maxPriceRange: int, maxSlippage: long) → int

Examples
Example 1
logs = [["3","A","103","100","2"],["1","A","100","100","5"],["2","B","90","90","1"],["2","A","101","100","2"]]
targetVenue = "A"
maxPriceRange = 3
maxSlippage = 8
return = 3
After filtering and sorting, all three A rows have range 3 and total slippage 8.
Example 2
logs = [["1","B","10","10","1"]]
targetVenue = "A"
maxPriceRange = 0
maxSlippage = 0
return = 0
No row belongs to the target venue.

Constraints
0 <= logs.length <= 500.
Prices, volume, and bounds are nonnegative integers.
Slippage totals fit in a 64-bit signed integer.

Parse the strings to integers first. Filter to the target venue, then use a stable sort on timestamp so equal timestamps keep input order. Precompute each row's cost as abs(fill - benchmark) * volume, and use a long for sums. Now run two pointers. Expand right, add the cost to a running sum, and track the min and max fill price in the window with two monotonic deques. While range exceeds maxPriceRange or the sum exceeds maxSlippage, shrink left. Both conditions only get easier as the window shrinks, so the window is monotone and this works. The common pitfall is skipping the stable sort, or using int for slippage and overflowing. With n up to 500, even an O(n^2) scan with running min and max passes, so don't over-engineer it. If you freeze mid-OA, StealthCoder is the hedge that gives you the full approach live.

Drill it cold or hedge it with StealthCoder. Either way, don't walk into the OA hoping you remember the trick.