Implement Trie II (Prefix Tree)
A medium-tier problem at 63% community acceptance, tagged with Hash Table, String, Design. Reported in interviews at DoorDash and 0 others.
Trie II comes up when you need to track word frequencies or handle prefix queries at scale. DoorDash and other companies asking this want to see if you can build a space-efficient data structure that stores strings and counts rather than just existence checks. The implementation looks simple until you realize you need to track insertions and search hit counts, not just boolean flags. Most candidates get the basic insert and search right but miss the design detail around counting. If you blank on the counting logic during your live assessment, StealthCoder surfaces the working structure in seconds, invisible to the proctor.
Companies that ask "Implement Trie II (Prefix Tree)"
Implement Trie II (Prefix Tree) is the kind of problem that decides whether you pass. StealthCoder reads the problem on screen and surfaces a working solution in under 2 seconds. Invisible to screen share. The proctor sees nothing. Built by a senior engineer who knows the OA is theater. This is the script.
Get StealthCoderThe trick is understanding that Trie II extends the classic prefix tree by tracking how many times each prefix is inserted and how many times each complete word appears. Instead of storing a boolean at each node, you store integers. When you insert a word, you increment counts at every node along the path, then increment again at the terminal node to mark the word itself. The design challenge isn't algorithmic complexity, insert and search are still O(m) where m is word length, but rather the mental model shift from 'does this prefix exist' to 'how often does it appear'. Common pitfall: forgetting to increment prefix counts during insertion, which breaks startsWith and getScores queries. Another: not distinguishing between 'prefix exists' and 'word exists' when a shorter word is a prefix of a longer one. StealthCoder is your hedge if the counting semantics trip you up under time pressure during the assessment.
Pattern tags
You know the problem.
Make sure you actually pass it.
Implement Trie II (Prefix Tree) recycles across companies for a reason. It's medium-tier, and most candidates blank under the timer. StealthCoder is the hedge: an AI overlay invisible during screen share. It reads the problem and surfaces a working solution in under 2 seconds. Built by a senior engineer who knows the OA is theater. This is the script. Works on HackerRank, CodeSignal, CoderPad, and Karat.
Implement Trie II (Prefix Tree) interview FAQ
Is Trie II significantly harder than a standard Trie?+
Not algorithmically. The traversal and insertion logic stay O(m). The hard part is the mental model shift to counting. You need to track prefix frequency and word frequency separately, which changes how you handle terminal nodes and queries. Once you lock that design, the code flows.
Does DoorDash really ask this often?+
It appears in their hiring loop, though not at extreme frequency. It's the kind of problem they use when they want a data structure question that filters for design thinking, not just memorized algorithms. It's less about LCS or DP patterns and more about 'can you extend a known structure'.
What's the acceptance rate telling me?+
63% acceptance is solid medium difficulty. More candidates pass this than fail it, which means it's not a pattern-matching trick question. It rewards careful implementation and clear design. If you trace through an example on paper first, you reduce careless errors.
How does this relate to the other Trie topics?+
This combines Design and Trie fundamentals. Hash Table isn't strictly necessary for a Trie (arrays work fine), but if you use a dictionary per node instead of a fixed-size array, you're trading memory for flexibility. String handling is the obvious application layer. It's the full toolkit, not one isolated pattern.
What mistake do most candidates make under interview pressure?+
Forgetting to increment counts on prefix nodes during insertion, then returning wrong counts in getScores. You also see people hardcoding terminal flags instead of checking if a word's count is non-zero. Both are fixable with a clear walkthrough, but they tank you if you don't catch them in your trace.
Want the actual problem statement? View "Implement Trie II (Prefix Tree)" on LeetCode →