Knowledge induction of common basic concepts in relational database

Statement: My article is only the main writing, minor details too much, induction is meaningless, at the same time induction is mainly to their own view,

And basically are some of their own memory and understanding of the instant writing, not necessarily right and full (but most of the concepts need to understand), please watch forgive me!

Database Design Chapter

1. Paradigm

A.1 Paradigm, atomicity, that is, the column is not divided

B.2 paradigm, completely dependent, that there is a primary key unique distinction

C.3 paradigm, cannot pass dependency, that is, table cannot have other table's non-primary key information

2. Model

A. Conceptual model, ER diagram, etc.

B. Logical model, which is the logical table

C. Physical model, that is, physical table generation

Transaction

1. Four characteristics,

A. Atoms, either. Either..

B. Isolation, unaffected

C. Consistency, consistency

D. Permanent, save

2. Isolation, read the question

A. Dirty read, READ UNCOMMITTED

B. Non-repeatable read, read to commit (update)---seems to be MSSQL default

C. Phantom (Virtual read, can be reread), read bar number (Delete|insert)---seems to be MySQL default

D. Serial, this does not have to say more.

3. Lock, write problem (missing update), concurrency control

A. Pessimistic lock (Database layer): Read lock (Shared lock) can not be modified after adding to the table---easy to deadlock; write lock (Exclusive lock) cannot add additional locks after adding to table

B. Optimistic lock (non-database): such as timestamp

4. Transaction passing, that is, nested calls, that is, whether you can not open a new transaction

5. Mode of Use, database, coding, declarative

6. Distributed, this almost forget, is the scope of the transaction and 2pc (this is OK and then action)

Connection pool, this is also the last bar, in fact, it is said that the design mode of the enjoyment of the meta-model of practical, generally written by others, the direct use is

The index, the aggregation, is physical, the non-aggregation is logical, and the joint index is multiple;

There is what SQL writing, left and right connection and so on, even if it is not what difficult to calculate the East four (not in the optimization level).

As to what security-related, database recovery and other DBA-related operations, and nosql& Big data is more out of this range, streamline induction.

There is a database of master-slave replication, read and write separation, data segmentation, clustering and so on are not on this basis, there is a chance to start another MySQL architecture optimization chapter.

There are still some said, such as read-write separation with the master-slave replication method to achieve. Alas, said more.

Ps

Writing this is mainly an interview this morning (again fooled once). The phone said that as long as you can, go to say we now only a simple development (really doubt whether the resume is written a problem?) ）

15K is already a manager level or something. To! Horizontal rotten (There are reading questions, more than half of the Runat=server user control of the memory of the bad problem!) ) also: Oh ~

The other half of his written test is SQL, and suddenly feel that there is a database is not covered in the basis of the summary ...

Yes! It seems that there is a network, and then write it.

Knowledge induction of common basic concepts in relational database