Cell-level response federal active calculation method
By setting nodes in multiple databases and forming active computing nodes using control data columns, cell-level response and federated active computing are realized, which solves the problems of complex logic, high computing cost and inability to handle multiple databases in the prior art, and realizes the effect of automated triggering and reducing computing frequency.
Patent Information
- Application Number
- CN202510166905.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-14
- Publication Date
- 2025-06-24
AI Technical Summary
The existing active computing methods have problems such as complex logic, high computational cost and inability to handle multiple databases when implementing cell-level response and federated active computing.
A cell-level response federal active computing method is adopted, by setting nodes in multiple databases, using control data columns to form active computing nodes, realizing an automated triggering and caching mechanism, and having cell-level judgment capabilities.
It realizes automatic triggering of active computing, reduces the frequency of computational triggering, has federal active computing capabilities, and can effectively handle the computing logic of multiple databases.
Smart Images
Figure CN120196637A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of active computing, and particularly to a cell-level response federated active computing method. Background Art
[0002] The methods for implementing active computing logic in the industry are usually materialized views + triggers / timed refresh / event-driven. The main disadvantages of this solution are 1. The logic of the materialized view also needs to be clearly known to the developers of triggers / event-driven. In addition to the trouble of developing the logic, it actually means maintaining two pieces of logic, and there is inconsistency in the supporting modification (for example, after the logic of the materialized view is modified, the trigger is forgotten to be modified correspondingly / the modification is incorrect, resulting in the incorrect active computing of the materialized view).
[0003] 2. In triggers / event-driven, cell-level triggering of active computing cannot be configured. This means that the materialized view is often triggered for unnecessary computing due to changes in the data of the physical table. It consumes huge computing costs.
[0004] The materialized view and triggers / event-driven can only be established on a single database, and are powerless when the computing logic involves multiple databases. Summary of the Invention
[0005] A cell-level response federated active computing method includes the following steps: Step 1, for a first node in a first database that meets the desired data requirements, the first node contains control data columns; Step 2, form a second node in a second database that contains the control data columns, and all data in the second database is contained in the second node; Step 3, form a third node for active computing according to the first node and the second node.
[0006] Preferably, the triggering conditions for the active computing of the third node include: Condition 1, the conditions of the first node are met; Condition 2, the conditions of the second node are met.
[0007] Preferably, the results of the active computing of the third node are in the active computing cache.
[0008] Preferably, the active computing involves cell-level judgment.
[0009] Therefore, the present invention does not require additional configuration of triggers / event-driven. When the computing node is configured for active computing, the computing results of the node are automatically cached, and it automatically senses changes in upstream data and decides whether to trigger computing. The automatic trigger of the active computing function has the cell-level judgment ability. Any computing logic can achieve column-level response judgment. When the computing logic contains conditional judgment, it can decide whether to trigger active computing based on the recorded content. For example, the computing logic is: select a, b, c from Table A where c > 10; Only when the changed data in the database table A is such that the data content of any column of a, b, c changes, and there is a record in the changed row where the value of column c is greater than 10: (a || b || c) && (c > 10). Can it possibly trigger active computing. During federated computing, it still has the ability of active computing.
[0010] The following effects are achieved: The ability to automate the trigger of active computing. There is no need to write additional trigger logic similar to triggers.
[0011] The ability to notify of cell-level judgment of data content changes that trigger active computing. It can effectively reduce the trigger frequency of active computing.
[0012] Federated active computing ability, that is, active computing can involve multiple databases. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] Figure 1 It is a schematic diagram of a cell-level response federated active computing method of the present invention. DETAILED DESCRIPTION OF THE INVENTION
[0014] Term Definition: CDC: Database data change capture technology. It can obtain the recorded content before and after the database change.
[0015] Node: A piece of fetching SQL code or a piece of computing logic; Federation: Simultaneously read or process data from multiple databases (or different types); Example illustration of cell-level response federated active computing ability: DB1 (Database 1): Define node nodeA: select a, b, c from Table A where c > 10; DB2 (Database 2): Define node nodeB: select * from TableB; (There is also a column c in tableB) Define a new node nodeC based on nodeA and nodeB and configure it for proactive calculation: Define node NodeC: / / Query the values in nodeB where the value of column column4 is "SDM" or the value of column5 is "USAW", and return the three columns c, column4, and column5 dfB = nodeB[c, column4, colunm5].query(column4 = "SDM" or column5 = "USAW"); / / Perform a join calculation between nodeA and nodeB, with the join key being column c nodeA.merge(dfB, on = nodeA[c]); NodeC will trigger proactive calculation only when one of the following two conditions is met: Data changes occur in tableA of DB1. The changed content meets the condition that any one of the data contents in columns a, b, and c changes, and there is a record in the changed row where the value of column c is greater than 10 (both c > 10 before and after the change are satisfied). Data changes occur in tableB of DB2. The changed content meets the condition that any one of the data contents in columns c, column4, and column5 changes, and the value of column4 is "SDM" or the value of column5 is "USAW" (both column4 = "SDM" || column5 = "USAW" before and after the change are satisfied). The proactive calculation trigger logic expression is expressed as: ((a || b || c) && (c > 10)) || ((c || column4 || column5) && ((column4 = "SDM") || (column5 = "USAW")) The engineering implementation is as Figure 1 shown: Federated proactive calculation is the calculation logic that requires proactive calculation. Such as nodeC configured for proactive calculation above.
[0016] The driving decision is based on the calculation logic (nodeA, nodeB, nodeC above), and automatically determines whether to trigger proactive calculation based on data changes in DB1 and DB2.
[0017] The proactive calculation cache is the calculation result of proactive calculation.
[0018] A cell-level response federated proactive calculation method includes the following steps: Step 1, for the first node in the first database that meets the desired data requirements, the first node contains control data columns; Step 2, form a second node in the second database that contains the control data columns, and all data in the second database is contained in this second node; Step 3, form a third node for proactive calculation based on the first node and the second node.
[0019] Preferably, the triggering conditions for the active calculation of the third node include: Condition 1, the condition of the first node is satisfied; Condition 2, the condition of the second node is satisfied.
[0020] Preferably, the result of the active calculation of the third node is in the active calculation cache.
[0021] The preferred embodiments of the present invention are exemplarily given above with the aid of the drawings. However, those skilled in the art should understand that without exceeding the protection scope of the appended claims, those skilled in the art can form new technical solutions by adding features, combining features, etc.
Claims
1. A cell-level response federation active computing method, characterized in that: The steps include: Step 1, for a first node in a first database, which is set to meet expected data requirements, the first node includes a control data column; Step 2, forming a second node in a second database including the control data column, wherein the second node contains all the data in the second database; Step 3: forming a third node for active computing according to the first node and the second node.
2. The cell-level response federation active computing method according to claim 1, characterized in that: The triggering conditions for the active calculation of the third node include: Condition 1, the condition of the first node is met; Condition 2, the condition of the second node is met.
3. The cell-level response federation active computing method according to claim 2, characterized in that: The result of the active calculation by the third node is in the active calculation cache.
4. The cell-level response federation active computing method according to claim 3, characterized in that: The active calculation involves cell-level judgment.