[hard copy] core differences among dirty reading, non repeatable reading and unreal reading scenarios

I think dirty reading can be read alone , Unrepeatable and unreal reading can compare memory

Dirty reading

• scene
  A Transaction modified data , At the moment B The transaction reads this data , And use this data to do logical processing ; then A The transaction returns the data just submitted for some reason , however B The transaction doesn't know about this. It commits the normal transaction ;

• concerns
  One of the biggest problems caused by dirty reading is dirty writing

It can't be read repeatedly

• scene
  A When the connection starts to query data 1, But the data queried later is 2, Because halfway through B The connection changes the data to 2, At this time, the data of the two queries are inconsistent

• concerns
  B The connection will A The number of connections that can be queried has been modify

Fantasy reading

• scene
  The number of data transactions in two queries of a transaction is inconsistent , for example A The transaction queries several columns (Row) data , And another one. B The transaction inserts new columns of data at this time , The previous A The transaction is in the next query , You'll find several columns of data that it didn't have before .

• concerns
  B Transaction in A Within the data range that the transaction can query newly added Data

contrast

It can't be read repeatedly The focus is on data modification , Fantasy reading The focus is on the addition of data