Store display labels, units, owners, and other metadata on tables and columns as annotations that any application or tool can read.

Decide what happens to child rows when their parent is deleted: refuse the delete, cascade it, or keep the children and clear the reference.

Check constraints at commit instead of after each statement, so rows that reference each other or swap values can be saved together.

Add columns that SELECT * and INSERT without a column list ignore, so existing code keeps working, and reorder columns along the way.

Define computed values such as a fare with tax once in the table with virtual columns, which take no space and can be indexed.

Generate surrogate keys automatically with identity columns, choose between ALWAYS and BY DEFAULT, and restart after loading data.

Compute each value's share of the total, such as revenue by payment method, overall or within groups, in one analytic function.

Get the value from any position in a window, such as the second hire or the second most recent row, counting from the first or the last.

Get the value from the last row of each group with LAST_VALUE, and avoid the default frame that makes it return the current row instead.

Test whether differences between groups are real with t-tests, ANOVA, and chi-squared tests run directly in SQL with the STATS_ functions.

Show the value from the first row of each group on every row with FIRST_VALUE, and compare each row with its group's starting value.

Count distinct integers exactly with bitmaps that can be stored per group and merged, so distinct counts over any period stay correct.

Number the rows of each group with ROW_NUMBER for top-N-per-group queries and deduplication, and make the numbering repeatable.

Answer top-N questions such as the busiest routes or best-selling cabins quickly with APPROX_COUNT, APPROX_SUM, and APPROX_RANK.

Estimate medians and percentiles of very large data quickly, make the result repeatable with DETERMINISTIC, and check its accuracy.

Estimate the number of distinct values in huge tables quickly, and store mergeable sketches so distinct counts add up correctly.

Detect whether data changed or differs between databases by comparing order-independent checksums of column values over groups.

Combine flag integers bit by bit over a group to find options common to all rows or available in any, and decode them with BITAND.