Top Performing Models

An easy SQL interview practice problem on DataDriven. Write the query and run it against a real database, with instant feedback.

Domain: SQL
Difficulty: easy
Seniority: L3

Problem

Which models in our ml_models registry have an accuracy of 0.90 or above? Return every field for each one, highest accuracy first.

The models that actually perform.

Tables

ml_models

Columns: model_id INTEGER, mdl_name TEXT, version TEXT, accuracy REAL, status TEXT, train_at TEXT, framework TEXT

ml_models: 10 of 200 rows
model_id	mdl_name	version	accuracy	status	train_at	framework
123	rec_engine	v1.1	0.72	training	2026-02-02 01:00:00	TensorFlow
146	churn_pred	v2.0	0.89	archived	2026-03-03 02:00:00	XGBoost
169	spam_filter	3.0	0.61	failed	2026-04-04 03:00:00	sklearn
192	img_classify	v2.1-beta	0.78	validating	2026-05-05 04:00:00	pytorch
215	sentiment	1.0.0	0.95	Deployed	2026-06-06 05:00:00	tensorflow
238	demand_fcst	v1.0	0.67	deployed	NULL	PyTorch
261	price_opt	v1.1	0.84	Training	2026-08-08 07:00:00	TensorFlow
284	fraud_detect	v2.0	NULL	archived	2026-09-09 08:00:00	XGBoost
307	rec_engine	3.0	0.73	failed	2026-01-10 09:00:00	sklearn
330	churn_pred	v2.1-beta	0.9	validating	2026-02-11 10:00:00	pytorch

Expected output

First 5 of 44 rows
model_id	mdl_name	version	accuracy	status	train_at	framework
951	sentiment	v1.1	0.99	training	2026-02-10 13:00:00	TensorFlow
1986	churn_pred	v2.1-beta	0.99	validating	2026-02-27 10:00:00	pytorch
10002437	sentiment	v1.1	0.99	training	2023-01-10 13:00:00	TensorFlow
2285	price_opt	1.0.0	0.95	Deployed	2026-06-12 23:00:00	tensorflow
2400	img_classify	v2.1-beta	0.9	validating	2026-02-17 04:00:00	pytorch

Worked solution and explanation

What this is really asking

This is a boundary-reading question dressed up as a model registry report. Nobody struggles with the filter itself. What separates candidates is whether 'or above' reaches the query as >=. Several models on this seed score exactly 0.90, including model_id 2400 in the preview. A query using accuracy > 0.90 runs cleanly, returns plausible rows and quietly loses every one of them. The other half is the order: the report is read strongest first, and the right rows in the wrong order still miss.

Step 1: Keep the 0.90 models in

Write the floor as accuracy >= 0.90. Those scores are stored as a plain 0.9, so the boundary is real data and not a floating-point accident. A strict comparison simply drops them.

Step 2: Return the whole model record

'Every field' means all seven columns of ml_models. Listing them by name returns the same row as SELECT *, and the output shape stays fixed if someone later adds a column to the registry.

Step 3: Put the strongest models on top

Sort by accuracy DESC. Several models share 0.99 and several share 0.90. The question says nothing about how equal scores are ordered, so any order among them is correct and no second sort key is owed.

Models at or above 0.90, strongest first

SELECT
  model_id,
  mdl_name,
  version,
  accuracy,
  status,
  train_at,
  framework
FROM ml_models
WHERE accuracy >= 0.90
ORDER BY accuracy DESC

Strict comparison: WHERE accuracy > 0.90 reads naturally and runs without complaint, but every model sitting exactly on 0.90 disappears, model 2400 among them.
Inclusive comparison: WHERE accuracy >= 0.90 keeps the boundary models, which is what 'or above' means to whoever reads the report.

Common follow-up questions

If the ask became 'above 0.90', which models leave the result? (Tests reading the boundary both ways: the models scoring exactly 0.90 drop out.)
How would you return only the best-scoring version of each `mdl_name`? (Moves from a plain filter to picking one row per group.)
What breaks if `accuracy` were stored as text such as '0.9'? (Text compares character by character, so the threshold needs a numeric conversion first.)