Ask2SQL is a free AI-powered tool that converts natural language questions into SQL queries. Type what you want in plain English and get an executable SQL statement for MySQL, PostgreSQL, BigQuery, Snowflake, or SQLite instantly.

How do I convert English to SQL?

Simply type your data question in plain English (e.g., 'Get the top 10 customers by total order amount'), select your database dialect, and click Generate. Ask2SQL's AI will produce the correct SQL query with proper syntax for your chosen database.

What SQL dialects are supported?

Ask2SQL supports MySQL, PostgreSQL, Google BigQuery, Snowflake, and SQLite. Each dialect generates syntax-specific SQL, including BigQuery's QUALIFY clause, PostgreSQL's ILIKE, and MySQL's LIMIT syntax.

Yes. You get 10 free SQL queries per day with no signup required. Pro plans start at $9/month for unlimited queries, priority processing, and query history.

What is schema-aware SQL generation?

Schema-aware generation lets you paste your CREATE TABLE definitions so the AI generates queries using your exact table and column names. This eliminates guessing and produces ready-to-run SQL without manual fixes.

SQL Date Functions Cheat Sheet: PostgreSQL, MySQL, BigQuery, SQL Server

Quick Answer

Truncate to a period: DATE_TRUNC('month', d) in PostgreSQL and BigQuery; MySQL has no equivalent — use DATE_FORMAT(d, '%Y-%m-01'); SQL Server 2022+ has DATETRUNC(month, d).
Add or subtract time: PostgreSQL uses interval arithmetic (d - INTERVAL '1 month'), the other three use functions (DATE_ADD, DATE_SUB, DATEADD).
End of month: LAST_DAY(d) covers MySQL and BigQuery; PostgreSQL needs DATE_TRUNC('month', d) + INTERVAL '1 month - 1 day'; SQL Server uses EOMONTH(d).

The Master Table

Task	PostgreSQL	MySQL 8	BigQuery	SQL Server
Now (timestamp)	`NOW()`	`NOW()`	`CURRENT_TIMESTAMP()`	`SYSDATETIME()`
Today (date)	`CURRENT_DATE`	`CURDATE()`	`CURRENT_DATE()`	`CAST(GETDATE() AS DATE)`
Truncate to month	`DATE_TRUNC('month', d)`	`DATE_FORMAT(d, '%Y-%m-01')`	`DATE_TRUNC(d, MONTH)`	`DATETRUNC(month, d)` (2022+)
Truncate to week	`DATE_TRUNC('week', d)` (Monday)	`DATE_SUB(d, INTERVAL WEEKDAY(d) DAY)`	`DATE_TRUNC(d, WEEK)` (Sunday)	`DATETRUNC(week, d)` (Sunday)
Add N days	`d + INTERVAL '7 days'` or `d + 7`	`DATE_ADD(d, INTERVAL 7 DAY)`	`DATE_ADD(d, INTERVAL 7 DAY)`	`DATEADD(day, 7, d)`
Subtract 1 month	`d - INTERVAL '1 month'`	`DATE_SUB(d, INTERVAL 1 MONTH)`	`DATE_SUB(d, INTERVAL 1 MONTH)`	`DATEADD(month, -1, d)`
Difference in days	`d2 - d1` (integer)	`DATEDIFF(d2, d1)`	`DATE_DIFF(d2, d1, DAY)`	`DATEDIFF(day, d1, d2)`
Extract year	`EXTRACT(YEAR FROM d)`	`YEAR(d)`	`EXTRACT(YEAR FROM d)`	`YEAR(d)`
Format as text	`TO_CHAR(d, 'YYYY-MM')`	`DATE_FORMAT(d, '%Y-%m')`	`FORMAT_DATE('%Y-%m', d)`	`FORMAT(d, 'yyyy-MM')`
Parse a string	`TO_DATE(s, 'YYYY-MM-DD')`	`STR_TO_DATE(s, '%Y-%m-%d')`	`PARSE_DATE('%Y-%m-%d', s)`	`TRY_PARSE(s AS DATE)`
End of month	`DATE_TRUNC('month', d) + INTERVAL '1 month - 1 day'`	`LAST_DAY(d)`	`LAST_DAY(d)`	`EOMONTH(d)`
Age in years	`EXTRACT(YEAR FROM AGE(now, d))`	`TIMESTAMPDIFF(YEAR, d, NOW())`	`DATE_DIFF(CURRENT_DATE(), d, YEAR)`	`DATEDIFF(year, d, GETDATE())`

Grouping by Week: The ISO Trap

"Revenue by week" is where date code goes wrong most often, because the four engines define week boundaries and week numbers differently. PostgreSQL's DATE_TRUNC('week', d) starts weeks on Monday; BigQuery and SQL Server start them on Sunday. If you need ISO week numbers (Monday-based, week 1 contains the first Thursday), MySQL gives them via DATE_FORMAT(d, '%x-W%v') — note the lowercase %x%v for ISO; uppercase %X%V is the Sunday-based variant. Mixing those two formats for year and week is a classic bug that splits one ISO week across two grouping buckets every January.

Adding Months: End-of-Month Drift

DATE '2026-01-31' + INTERVAL '1 month' in PostgreSQL returns 2026-02-28 — the last valid day. MySQL's DATE_ADD('2026-01-31', INTERVAL 1 MONTH) returns the same. But SQL Server's DATEADD(month, 1, '2026-01-31') also returns 2026-02-28, while DATEADD(month, 1, '2026-01-30') returns 2026-02-28 too (it clamps, never overflows). The trap runs the other direction as well: repeated "+1 month" on a clamped date (Jan 31 → Feb 28 → Mar 28) drifts away from the true month-end. For subscription and billing queries, always compute from an anchor date like the first of month, not by iterating.

Timezones

PostgreSQL stores timestamptz in UTC and renders in the session timezone: created_at AT TIME ZONE 'America/New_York' converts explicitly. MySQL's TIMESTAMP auto-converts using the session time_zone, but DATETIME does not convert at all — half the "my dates are off by an hour" reports trace to this. For date-bucket reporting, the safe pattern in every engine is: store UTC, convert once at query time, and group by the converted value — never group by UTC day and label it with a local timezone.

Filling Date Gaps

Date functions get you the axis; the missing-days problem is separate — generating a complete date spine and LEFT JOINing your data onto it. We covered that with dialect-specific recipes here: Find missing dates in SQL.

Try It Yourself

Describe the date logic in plain English — "monthly revenue for the last 6 months in MySQL" — and our free SQL generator writes the dialect-correct functions for you.