• Topic
  • Discussion
  • UdaWikiWeb.JDBCSensitivityToChange(Last) -- Owiki? , 2016-08-19 15:00:29 Edit owiki 2016-08-19 11:00:29

    Additional Reading

    Sensitivity To Changes In Underlying Database

    It is extremely important to application developers and end-users alike to understand the degree to which the Resultsets presented to them by a JDBC solution are actually sensitive to underlying changes in the source database. JDBC 1.1 not only fails to provide you with Bi-directional Resultset Scrolling, it also presents what is basically a snapshot of the data in your database at the time a JDBC query is executed. This has the effect of increasing Multi-User JDBC solution development complexity or limiting the functionality and usability of JDBC by end-users.

    Sensitive to changes in underlying database takes many forms, this includes: Static, KeySet, Dynamic, and Mixed modes of sensitivity.

    Static

    Same as basic JDBC, records scrolling occurs over a database snapshot and is insensitive to underlying change by other users

    KeySet

    JDBC resultset records scroll over a set of record identifiers uniquely identifying records in the underlying database, this type of scrolling is sensitive to changes is those records with identifiers at the time of query execution. This form of scrolling is insensitive to record record additions or deletions.

    Dynamic

    JDBC resultset records scroll over a set of record identifiers uniquely identifying records in the underlying database, these unique identifiers are recreated before each RowSet traversal (collection of resultset records used as scrolling marker or sliding window or Cursor), rather than once at query execution time. This type of scrolling is sensitive to all changes in the underlying database but may introduce a performance penalties depending on the size of RowSets and available network bandwidth.

    Mixed

    JDBC resultset records scroll over a set of record identifiers uniquely identifying records in the underlying database, these unique identifiers are created to a limited size (known as the KeySet Size) at query execution time, only when RowSet traversal goes beyond the existing set of unique row identifiers is another collection of unique identifiers assembled. This type of scrolling is sensitive to all changes in the underlying database, but insensitive to Additions or Deletions affecting records in the current RowSet scrolling across a current KeySet, once KeySet boundaries are crossed Insertions or Deletions are recognized. This mode of sensitivity provides increased performance and the expense of reduced sensitivity.


    Referenced by...