Oracle SQL

Store each month of a large table in its own partition with range and interval partitioning, and read only the partitions a query needs.

Store rows inside the primary key index with ORGANIZATION INDEX for fast key lookups and range scans, and know when an IOT fits.

Query CSV and other files on the database server as tables without loading them, with an external table or an inline EXTERNAL clause.

Create tables that exist only in your session and vanish by themselves, without committing the transaction, and compare them with GTTs.

Bring back a dropped table from the recycle bin, restore it under a new name, and rewind a table's rows after a wrong DELETE or UPDATE.

Make installation scripts safe to run twice with CREATE ... IF NOT EXISTS and DROP ... IF EXISTS, without PL/SQL checks around DDL.

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.