HFSQL

TRANSACTIONS: ACID A transaction is a set of indissociable oper- ations: either all the operations of the transaction are performed, or none is per- formed. Transaction management is the best way to ensure the integrity of a set of indisso- ciable write operations performed on HFSQL tables. A transaction is used to make sure that the updates performed on one or more tables have reached completion successfully. HFSQL supports all the types of transac- tions, and therefore meet the ACID criteria (ACID is the acronym for Atomicity, Con- sistency, Isolation and Durability). HFSQL Client/Server offers 4 isolation modes for transactions. • Unvalidated data (READ UNCOMMIT- TED) • Validated data (READ COMMITTED) • Instant transaction photography (RE- PEATABLE READ) • Serializable transaction (SERIALIZABLE). REPLICATION HFSQL proposes 4 types of replication: • HFSQL server replication • replication between heterogeneous data- bases, a HFSQL and Oracle database for example • replication with mobile devices (iOS, An- droid, ...) • offline replication, without permanent link A replication is easily defined via the repli- cation wizard, or via programming. AUTOMATIC ROW LOCKING HFSQL supports locks at the table level and at the row level. Support for locks at the row level ensures better access security. This management is automatic. AUTOMATIC RECONNECTION This feature automatically manages dis- connections happening between the client and the server. Usually, this problem occurs with hardware whose connection with the server is not al- ways on: mobile devices (Wi-Fi, 3G, 4G, ...) notably. This can also happen on regular wired net- works. When the application is reconnected, the application needs to resume at the point where the connection was interrupted, and make sure the buffers are accurate. With the automatic resuming of the con- nection, all the buffers and positions are stored and reassigned. The application can resume without any error, as if the connection was never inter- rupted. You can also manage the connection inter- ruptions by programming, or execute addi- tional processes if you want. STORED PROCEDURES (UDF) The stored procedures (sometime called UDF) are used to simplify the development and the maintenance of your applications by factorizing the code. Indeed, when the method for calculating a result or a business rule found in a stored procedure evolves, all you have to do is modify the stored procedure on the server without performing any modification in the deployed applications. A stored procedure is also used to limit the number of back and forth processes be- tween the client computer and the server, and therefore increasing the speed of processes. The same stored procedure can be shared among several applications. Stored procedures are programmed using WLanguage, and therefore benefit from the feature richness and ease of use of the 5GL. The creation of a stored procedure from the WINDEV or WEBDEV environment is really easy. TRIGGERS A trigger allows you to trigger a stored procedure before or after an event on a database table: for example when deleting a row, or after modifying a row. A trigger brings a lot of security. The trigger will get triggered regardless of the application or component that ac- cesses the database and that performs the defined operation, without the applica- tion’s developer having to program any- thing. Server triggers, as their name indicate, are run directly on the server. The right to create a trigger is defined via the database’s rights. The WINDEV environment indicates to the developer the presence of these triggers. The triggers are viewed in the analysis (the data description). INTEGRATED SCHEDULER (SCHEDULED TASKS) HFSQL has an integrated scheduler that lets you define and configure scheduled tasks. A task consists not only in executing a stored procedure (UDF), but also in trigger- ing a backup, or in forcing a performance optimization request. Graphical and user-friendly inter- face of the scheduler (scheduled tasks) (scheduled tasks) The definition is done in the Control center or via programming. You can create, add, modify, enable, disable tasks scheduled via programming, or from the administration tool, as long as you have the proper rights. The scheduler allows the DBA to program the automated execution of tasks on the server: it also allows you to create batch processes. Tasks can be run at a set date, and re- peated at regular intervals. LOGS The log is a special table where all the op- erations performed on one or more tables from a given time are automatically stored. The log contains the history of the logged tables: author, date and time, before/after value, application name, IP ... The following operations can be per- formed from a log: • restore the content of a logged table if the data is lost or destroyed • restore the content of a table up to a given date HFSQL • www.windev.com 8