How should I practice Video Cache Hit Rates from a CSV Log?

Start by reading the full statement, solve the visible examples, then compare the pattern against related Tubi phone screen problems in FastPrep.

Is Video Cache Hit Rates from a CSV Log associated with Tubi?

FastPrep tags this as a Tubi phone screen coding practice problem when the problem catalog has that company and stage context.

Video Cache Hit Rates from a CSV Log

Tubi● MediumFULLTIMEPHONE SCREEN

Problem statement

Source note: In a Tubi technical phone screen on CoderPad, the task was a log parser: a CSV file of CDN video requests, each a cache hit or miss, had to be read to find each video's cache hit rate, then report the highest rate that is not 1 and the lowest rate that is not 0, each with its log details. The video ID had to be pulled from URLs that came in two different formats. The report gives no column names, sample lines, URL formats or tie rules.

Practice choices: the file is passed as an array of lines with the header timestamp,status,url; the two video URL formats (a path form and a query form), the CDN host and the non-video requests are ours; requests for one ID count together across both formats; the details reported for each answer are the video ID with its hit and request counts, unreduced; ties go to the smallest ID in string order; an answer with no qualifying video is NONE; and the size limits are ours. All three examples are ours.

The judged task matches the report at about 70%.

A video CDN writes one CSV log line per request. You get the file as an array of lines, logLines. The first line is the header timestamp,status,url. Each other line has a request time, a cache status of hit or miss, and the requested url.

Only video requests count. A video URL comes in one of two formats, and both carry the video ID:

Path format: https://cdn.example.com/videos/<id>/<file>, for example https://cdn.example.com/videos/v1/seg1.ts has ID v1.
Query format: https://cdn.example.com/stream?vid=<id>&seg=<n>, for example https://cdn.example.com/stream?vid=v1&seg=2 also has ID v1.

Skip every other request (static files and health checks). Requests for the same ID count together, whichever format they use.

For each video, its cache hit rate is hits / total, where total is its number of requests and hits is how many of them have status hit. Return an array of two strings:

The video with the highest hit rate among videos whose rate is not 1.
The video with the lowest hit rate among videos whose rate is not 0.

Write each answer as "<id> <hits>/<total>" with that video's own counts, not reduced (for example "a7 2/4"). If several videos tie on the rate, choose the smallest ID in plain string order. If no video qualifies, that answer is "NONE".

Function

cacheHitRateExtremes(logLines: String[]) → String[]

Examples

Example 1

logLines = ["timestamp,status,url","1,hit,https://cdn.example.com/videos/v1/seg1.ts","2,miss,https://cdn.example.com/stream?vid=v1&seg=2","3,hit,https://cdn.example.com/videos/v2/seg1.ts","4,hit,https://cdn.example.com/stream?vid=v2&seg=2","5,miss,https://cdn.example.com/static/logo.png","6,hit,https://cdn.example.com/videos/v3/seg1.ts","7,miss,https://cdn.example.com/videos/v3/seg2.ts","8,miss,https://cdn.example.com/stream?vid=v3&seg=3","9,miss,https://cdn.example.com/videos/v4/seg1.ts"]return = ["v1 1/2","v3 1/3"]

The static line is not a video request, so it is skipped.
Both URL formats give video IDs. The counts are v1: 1 hit of 2, v2: 2 of 2, v3: 1 of 3, v4: 0 of 1.
Leaving out v2 (rate 1), the highest rate is v1 with 1/2.
Leaving out v4 (rate 0), the lowest rate is v3 with 1/3.

Example 2

logLines = ["timestamp,status,url","10,hit,https://cdn.example.com/stream?vid=b2&seg=1","11,miss,https://cdn.example.com/videos/b2/seg1.ts","12,hit,https://cdn.example.com/videos/a7/seg1.ts","13,miss,https://cdn.example.com/stream?vid=a7&seg=5","14,hit,https://cdn.example.com/videos/a7/seg9.ts","15,miss,https://cdn.example.com/stream?vid=a7&seg=7","16,hit,https://cdn.example.com/health","17,hit,https://cdn.example.com/videos/c3/seg1.ts","18,miss,https://cdn.example.com/videos/z9/seg1.ts","19,miss,https://cdn.example.com/stream?vid=z9&seg=2"]return = ["a7 2/4","a7 2/4"]

The counts are a7: 2 of 4, b2: 1 of 2, c3: 1 of 1, and z9: 0 of 2. The health check is skipped. a7 and b2 both have rate one half, which is the highest rate below 1 and also the lowest rate above 0. On a tie the smaller video ID wins, so both answers are a7, written with its own counts 2/4.

Example 3

logLines = ["timestamp,status,url","20,hit,https://cdn.example.com/videos/v6/seg1.ts","21,miss,https://cdn.example.com/static/style.css","22,hit,https://cdn.example.com/stream?vid=v5&seg=1","23,hit,https://cdn.example.com/videos/v5/seg2.ts"]return = ["NONE","v5 2/2"]

Every video has rate 1 (v5: 2 of 2, v6: 1 of 1), so no video has a rate below 1 and the first answer is "NONE". For the second answer both rates tie at 1, and "v5" is the smaller ID.

Constraints

1 <= logLines.length <= 501
logLines[0] is exactly "timestamp,status,url".
Every other line is timestamp,status,url, where timestamp is 1 to 10 digits and status is "hit" or "miss".
Every url has one of these four forms: https://cdn.example.com/videos/<id>/<file>, https://cdn.example.com/stream?vid=<id>&seg=<n>, https://cdn.example.com/static/<file>, or https://cdn.example.com/health.
<id> has 1 to 8 lowercase letters or digits, <file> has 1 to 12 lowercase letters, digits or dots, and <n> has 1 to 4 digits.

See Tubi hiring insights