Skip to content
UniCircle
Menu
Sign inJoin
EasyCodingLogging

Summarise an HTTP access log

Question NODE-COD-003

A Node.js server writes one line per request: method, path, status code and response time in milliseconds. Produce a short summary:

  1. For each status class that occurs, in the order 2xx, 3xx, 4xx, 5xx, print <class>: <count> requests, avg <avg> ms, where avg is the mean response time rounded down to an integer.
  2. Finally print slowest: <method> <path> <ms> ms for the slowest request (the earliest line if tied).

Constraints

  • 1 ≤ n ≤ 10^5
  • Status codes are between 200 and 599; response times are integers between 0 and 10^6.

Input and output

Input
Line 1: n Next n lines: METHOD PATH STATUS MS
Output
The summary lines described above.

Examples

Input 1

5
GET /api/colleges 200 35
POST /api/login 401 12
GET /api/colleges/7 200 48
GET /old 301 3
GET /api/report 500 950

Output

2xx: 2 requests, avg 41 ms
3xx: 1 requests, avg 3 ms
4xx: 1 requests, avg 12 ms
5xx: 1 requests, avg 950 ms
slowest: GET /api/report 950 ms

2xx: (35 + 48) / 2 = 41.5, rounded down to 41.

Input 2

2
GET /a 404 10
GET /b 404 10

Output

4xx: 2 requests, avg 10 ms
slowest: GET /a 10 ms

A tie for slowest goes to the earlier line.

Show hints
  • The class is the first digit of the status code.
  • Keep a running count and sum per class instead of storing every time.
Show the explanation

One pass: for each line, take the first digit of the status as the class and add to that class's count and sum; track the slowest request, replacing it only when a time is strictly greater (that keeps the earliest on ties).

Then print the classes in fixed order, skipping empty ones, with Math.floor(sum / count).

In production you would read the log as a stream (readline over fs.createReadStream) so a multi-gigabyte file never sits in memory, and you would usually report percentiles (p95, p99) rather than averages, because a few very slow requests are hidden by a mean.

Original UniCircle question. Last reviewed 22 Oct 2026.