You are on page 1of 4

DAY 1

EF CODD – creator of relational algebra – consists of 11 or 12 rules


An RDBMS should follow at least 8 rules.

ORDBMS
- Arrays
- Nested tables

Block Chaining (char /varchar)


In case of updating a data of varchar format with a value that is greater than its previous
value then a separate record in another block is created with a reference to its address to
its initial location

PCT-USED / PCT-FREE
DAY 2
ESC char in LIKE

Non Equi join


Instead of doing an EXCEPT you can do a RIGHT / LEFT join and do a filter on NULL

Use DBLINK to get data from tables in 2 different databases

ANY, ALL operator can be used with > OR < OR => OR <=
DAY 3
Flashback
Control file
Scn = system change no. (for every transac)

Autonomous transaction

DAY 4
Tablespace
Different tables spaces for index n tables n storing data temporary queries

Vacuum in postgres is like defragmentation

SAVEPOINT in context of Transaction Rollback and Commit.


Ability to roll back to the last query o the failed transaction

System level privileges


Object Level privileges

SYNONYM in postgres is like a permanent ALIAS


SEQUENCE is like an AUTO INCREMENTOR so… it might be useful for generating IDs
Different options are available like CYCLE (restart incrementing after a specific value)
and STOP(stop incrementing after a specific value)
CACHE is also an option available

MATERIALIZED VIEW is a physical version of a normal VIEW where the data generated by a VIEW
is stored physically permanently and not just the query of the VIEW.
REFRESH can be applied on the VIEW periodically using a batch job so the most recent version of
the query result is available.

PACKAGE is like an encapsulation for multiple functions somewhat like a CLASS.


The performance benefit is that if a function of a package is invoked then the whole
PACKAGE BODY i.e. all functions’ code is brought into memory. So if a function from the same
package another I/O operation is not needed to read it from the disk as it is already present in
memory.

Btree INDEX and Bitmap / GIN INDEX


Bitmap takes the median as the root of the tree so that the tree doesn’t become skewed. I Think

EXPLAIN….QUERY to get stats on the query runtime and reads and writes performed for the
query.

You might also like