pandas vs SQL: таблица соответствий

pandas и SQL решают одни задачи — разными синтаксисами. Понимая SQL, вы будете быстро читать pandas-код и наоборот.

Таблица соответствий

pandas	SQL
`df.head(10)`	`SELECT ... LIMIT 10`
`df[df['plan'] == 'pro']`	`WHERE plan = 'pro'`
`df.sort_values('amount', ascending=False)`	`ORDER BY amount DESC`
`df[['id','name']]`	`SELECT id, name`
`df.groupby('category').agg({'price': 'sum'})`	`GROUP BY category` + `SUM(price)`
`df.merge(users, on='user_id')`	`JOIN users ON ...`
`df.pivot_table(values='amount', index='category', aggfunc='sum')`	`CASE WHEN` + `GROUP BY`
`df['amount'].cumsum()`	`SUM() OVER (ORDER BY ...)`
`df['category'].value_counts()`	`GROUP BY category ORDER BY COUNT(*) DESC`
`df.drop_duplicates('user_id')`	`DISTINCT ON (user_id)`
`df.isna()`	`IS NULL`

python

# pandas:
df.groupby('category')['price'].mean().sort_values(ascending=False)

sql

-- SQL:
SELECT category, ROUND(AVG(price), 2) AS avg_price
FROM products
GROUP BY category
ORDER BY avg_price DESC;