Multi-source database retrieval method and system for lamp production quality traceability

Abstract

The present application relates to the technical field of database retrieval and production quality traceability, and discloses a multi-source database retrieval method and system for lamp production quality traceability, which comprises the following steps: locating an initial data set in a plurality of preset data sources according to an input production batch identifier; performing hierarchical classification processing on the initial data set to disassemble it into process data subsets corresponding to a plurality of production process nodes; performing time sequence synchronization processing based on a timestamp alignment rule for each process data subset to obtain synchronized process data subsets; performing conflict detection and resolution processing based on a preset set of physical constraint conditions on the synchronized process data subsets to obtain standardized process data subsets; and linking the standardized process data subsets according to an actual process flow sequence to form a production history chain covering the whole process from raw material delivery to finished product storage and outputting the same. The present application realizes accurate quality traceability for the whole process of OLED lamp production.

Description

Technical Field

[0001] This invention relates to the field of database retrieval and production quality traceability technology, specifically to a multi-source database retrieval method and system for traceability of lighting fixture production quality. Background Technology

[0002] With the rapid development of the OLED lighting industry, OLED lighting products are increasingly widely used in the display lighting field. Their production processes are complex and quality requirements are stringent, making the need for quality traceability throughout the entire production process increasingly urgent. Existing technologies have developed some product quality traceability systems, such as those that integrate and map data from multiple information systems like enterprise resource planning (ERP) systems and logistics systems through product traceability platforms. This enables the recording and monitoring of product flow information across the entire supply chain, from procurement and production to distribution. However, these existing traceability systems focus primarily on synchronizing business documents at the ERP level, such as production plans and inbound / outbound slips. They are insufficient in collecting and correlating dynamic process data from different data sources, such as real-time operating parameters and environmental parameters of equipment, during production execution. When multiple sources of data corresponding to the same production batch have discrepancies in the time dimension, existing systems struggle to accurately align time-series data from different sources on a unified timeline. This results in a lack of accurate time-series correspondence between equipment parameters, raw material information, and environmental data at the same process node, thus affecting the accuracy and completeness of quality traceability results at the process granularity level. Summary of the Invention

[0003] The purpose of this invention is to provide a multi-source database retrieval method and system for tracing the quality of lighting fixture production, so as to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a multi-source database retrieval method for traceability of lighting fixture production quality, comprising the following steps: S1. In response to the input production batch identifier, locate the corresponding initial data set in multiple preset data sources according to the production batch identifier; wherein, the multiple preset data sources include a first database storing lighting production plan information, a second database storing real-time operating parameters of lighting production equipment, a third database storing batch information of lighting production raw materials, and a fourth database storing lighting production environment parameters. S2. Perform hierarchical classification processing on the initial data set. Based on the production process links associated with the data, the initial data set is decomposed into process data subsets corresponding to multiple production process nodes. Each process data subset contains equipment operation parameter fragments, raw material batch association information, and environmental parameter records under the corresponding production process node, which identify the production batch. S3. For each process data subset, perform time synchronization processing based on timestamp alignment rules, and align the equipment operating parameter fragments, raw material batch association information and environmental parameters from different data sources within the process data subset on a unified time axis to obtain the aligned synchronized process data subset. S4. Perform conflict detection and resolution on the synchronous process data subset: Based on the preset set of physical constraints, traverse the data items in the synchronous process data subset, filter out data items that contradict the preset set of physical constraints, and extract two data items in the synchronous process data subset that are adjacent to the filtered data items at the process node and satisfy the preset set of physical constraints. Use the values ​​of the two data items to perform linear interpolation to fill in the filtered data items, and obtain a standardized process data subset; wherein, the preset set of physical constraints includes production process sequence constraints and equipment state mutual exclusion constraints; S5. Link the standardized process data subsets according to the actual process flow sequence in the lighting production process to form a complete production history chain covering the entire process of lighting products corresponding to the production batch identifier, from raw material input to finished product warehousing, and output the production history chain as the search result.

[0005] As a preferred embodiment of the present invention, the specific implementation of locating the corresponding initial data set among multiple preset data sources in S1 is as follows: Extract the product model code and production date code segment from the production batch identifier; The product model code is matched with the planned product model field in the first database and the raw material matching model field in the third database, and the production date code is matched with the operation log date field in the second database and the environmental record date field in the fourth database. For multiple successfully matched record entries, extract the data generation timestamps of each record entry, sort the record entries according to the chronological order of the data generation timestamps, and check whether the timestamp interval between adjacent record entries after sorting exceeds the preset sampling interval threshold. If it exceeds the threshold, insert null placeholders between adjacent record entries, and merge all record entries after inserting null placeholders into the initial data set.

[0006] As a preferred embodiment of the present invention, the specific implementation of S2, which involves decomposing the initial data set into process data subsets corresponding to multiple production process nodes based on the production process links associated with the data, is as follows: Extract the workstation equipment identification code carried in each data record in the initial dataset; The corresponding process node number is retrieved from the preset process mapping table based on the workstation equipment identification code; the process mapping table stores the correspondence between multiple workstation equipment identification codes and multiple process node numbers; Data records with the same process node number are grouped into the same process data subset; For data records that do not have workstation equipment identification codes, the equipment type code and operation time interval carried by the data record are extracted. Based on the equipment type code, the corresponding candidate process node set is matched from the preset equipment and process mapping table. Then, the operation time interval is compared with the standard operation time window of each process node in the candidate process node set. The candidate process node whose overlap time exceeds the preset overlap threshold is taken as the process node to which the data record belongs, and the data record is assigned to the process data subset corresponding to the process node.

[0007] As a preferred embodiment of the present invention, the timing synchronization processing based on timestamp alignment rules in S3 specifically includes: The data source with the highest sampling frequency is selected from the subset of process data as the baseline time axis; For other data sources within the process data subset, identify the original timestamp carried by each data record in the other data sources, find the reference time point on the reference time axis with the smallest time difference from the original timestamp, and associate the data record with the reference time point. Obtain the standard timestamp sequence of each reference time point on the reference time axis, as well as the time deviation value generated when each data record in other data sources is associated with the reference time point. Extract the deviation value with the highest frequency from all time deviation values ​​as the system deviation compensation amount of other data sources. After adding the system deviation compensation amount to the original timestamps of all data records in other data sources, re-execute the reference time point lookup and association operation, and package and mark all the re-associated data records. All the marked data records share the same time axis index value.

[0008] As a preferred embodiment of the present invention, before decomposing the initial data set into process data subsets corresponding to multiple production process nodes according to the production process links associated with the data in step S2, the following operations are performed: Extract the lamp type code from the production batch identifier, and retrieve the corresponding standard process sequence from the preset process template library according to the lamp type code; the standard process sequence contains multiple standard process nodes arranged according to the actual production sequence of the lamps, and each standard process node carries the standard operation time threshold and standard workstation equipment type set under the lamp type; Obtain the actual workstation equipment type identifier and actual operation time carried by each data record in the initial dataset; Match each actual workstation equipment type identifier with the set of standard workstation equipment types carried by each standard process node, and compare the actual operation time recorded in the data with the standard operation time threshold carried by the standard process node; When the actual workstation equipment type identifier is successfully matched and the actual operation time does not exceed the standard operation time threshold, the data record will be classified into the temporary classification set corresponding to the standard process node. When there are free data records in the initial data set that have not been assigned to any temporary classification set, extract the adjacent process data association markers carried by the free data records, locate the target temporary classification set that has a process adjacency relationship with the free data records from the already classified temporary classification sets based on the adjacent process data association markers, and merge the free data records into the target temporary classification set. All temporary classification sets are output as process data subsets corresponding to multiple production process nodes, according to the arrangement order of the corresponding standard process nodes in the standard process sequence.

[0009] As a preferred embodiment of the present invention, the specific implementation of forming the production history chain in S5 is as follows: According to the sequence of process flow, the data subsets of each standardized process are arranged into a data link sequence; For two adjacent standardized process data subsets in the data link sequence, extract the set of first station equipment numbers and the set of first process end timestamps carried by all data items in the first standardized process data subset, and extract the set of second station equipment numbers and the set of second process start timestamps carried by all data items in the second standardized process data subset. Perform an intersection operation on the set of equipment numbers for the first workstation and the set of equipment numbers for the second workstation to obtain a subset of shared workstation equipment numbers. Extract the first end timestamp corresponding to the subset of shared workstation equipment numbers from the set of first process end timestamps, and extract the second start timestamp corresponding to the subset of shared workstation equipment numbers from the set of second process start timestamps. Detect whether the time interval between the first end timestamp and the second start timestamp is less than the preset process transition time threshold. When the time interval is less than the threshold for process transition time, a link pointer based on the shared workstation equipment number subset is established between the previous standardized process data subset and the next standardized process data subset. When the time interval is greater than or equal to the threshold of the process transition time, the workstation equipment number pair with the smallest time interval is selected from the set of timestamps at the end of the first process and the set of timestamps at the beginning of the second process. Based on the workstation equipment number pair, an alternative link pointer is established between the data subset of the previous standardized process and the data subset of the next standardized process, and the alternative link pointer is marked as pending manual verification. Iterate through all adjacent relationships, establish all link pointers and alternative link pointers, and output the production history chain with link tags.

[0010] As a preferred embodiment of the present invention, after outputting the production history chain with link markers in step S5, the following operations are also performed: Traverse each link pointer and alternative link pointer in the production history chain, extract all workstation equipment number pairs corresponding to the alternative link pointers, and form a sequence of equipment number to be verified. Based on the sequence of equipment numbers to be verified, query the physical location coordinates and equipment type identifiers corresponding to each workstation equipment number from the preset workstation equipment topology diagram; Two workstation equipment numbers that are less than the preset adjacency distance threshold and have the same equipment type identifier are grouped into the same equipment association group. For alternative link pointers within the same device association group, the marking status of the alternative link pointers will be automatically switched from pending manual verification to confirmed status, and the switching record will be retained. For candidate link pointers that are not assigned to any device association group, maintain the status of pending manual verification and output a pending verification prompt message in the production history chain.

[0011] A multi-source database retrieval system for traceability of lighting fixture production quality, used to execute the multi-source database retrieval method for traceability of lighting fixture production quality described in this invention, including: The multi-source positioning module is used to respond to the input production batch identifier and locate the corresponding initial data set in multiple preset data sources according to the production batch identifier; wherein, the multiple preset data sources include a first database storing lighting production plan information, a second database storing real-time operating parameters of lighting production equipment, a third database storing batch information of lighting production raw materials, and a fourth database storing lighting production environment parameters. The hierarchical classification module is used to perform hierarchical classification processing on the initial data set. Based on the production process links associated with the data, the initial data set is broken down into process data subsets corresponding to multiple production process nodes. Each process data subset contains equipment operation parameter fragments, raw material batch association information, and environmental parameter records under the corresponding production process node, which identify the production batch. The timing synchronization module is used to perform timing synchronization processing based on timestamp alignment rules for each process data subset. It aligns the equipment operating parameter fragments, raw material batch association information and environmental parameters from different data sources within the process data subset on a unified time axis to obtain the aligned synchronized process data subset. The conflict resolution module is used to perform conflict detection and resolution processing on the synchronous process data subset: Based on a preset set of physical constraints, it traverses the data items in the synchronous process data subset, filters out data items that contradict the preset set of physical constraints, and extracts two data items in the synchronous process data subset that are adjacent to the filtered data items at the process node and satisfy the preset set of physical constraints. The values ​​of the two data items are used to perform linear interpolation to fill in the filtered data items, resulting in a standardized process data subset; wherein, the preset set of physical constraints includes production process sequence constraints and equipment state mutual exclusion constraints; The traceability chain construction module is used to link standardized process data subsets according to the actual process flow sequence in the lighting production process, forming a complete production traceability chain that covers the entire process of lighting products corresponding to the production batch identifier from raw material input to finished product warehousing, and outputs the production traceability chain as search results.

[0012] As a preferred embodiment of the present invention, the multi-source positioning module further integrates a data pre-inspection submodule and a data routing submodule: The data pre-inspection submodule is used to retrieve successfully matched record entries from each preset data source, traverse the data type identifier carried by each record entry, and classify the record entries into structured data items and unstructured data items according to the data type identifier. The structured data items are directly stored in the first buffer of the multi-source positioning module, and the unstructured data items are stored in the second buffer of the multi-source positioning module. The data routing submodule is used to periodically poll the storage status of the first and second caches. When the cumulative storage of structured data items in the first cache reaches a preset batch capacity threshold, all structured data items in the first cache are grouped according to the data generation timestamp of the record entries, generating a structured batch package and pushing it to the hierarchical classification module. When the cumulative storage of unstructured data items in the second cache reaches a preset batch capacity threshold, the key fields of the process steps carried by each unstructured data item are extracted, and all unstructured data items in the second cache are pre-classified according to the key fields of the process steps, generating an unstructured batch package with process step tags and pushing it to the hierarchical classification module.

[0013] As a preferred embodiment of the present invention, the hierarchical classification module further integrates a process node verification submodule and a data correction submodule: The process node verification submodule is used to traverse each process data subset after the hierarchical classification module groups data records with the same process node number into the same process data subset. It checks whether each process data subset contains a complete process start timestamp and process end timestamp. For target data records that are missing process start timestamps or process end timestamps, it extracts the process start timestamps and process end timestamps carried by other data records in the process data subset to which the target data record belongs. It selects the earliest time value from the extracted process start timestamps as the process start timestamp of the target data record, and selects the latest time value from the extracted process end timestamps as the process end timestamp of the target data record. The data correction submodule is used to extract the process node number of the process data subset to which the target data record belongs after the process node verification submodule completes the timestamp completion of the target data record. It then queries the standard operation time range corresponding to the process node number from the preset standard process time mapping table, checks whether the actual operation time between the process start timetamp and the process end timetamp of the target data record falls within the standard operation time range. If it does not fall within the range, the process end timetamp of the target data record is corrected to the sum of the process start timetamp and the standard operation time range, and a correction log record is generated.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention performs hierarchical classification processing on multi-source data according to the production process steps, decomposing the initial data set from different databases into process data subsets corresponding to multiple process nodes. Each process data subset contains complete fragments of equipment operating parameters, raw material batch association information, and environmental parameter records for that process node. This overcomes the data integration defects of existing traceability systems that only stay at the business document level, and realizes accurate location and traceability of OLED lighting product quality anomalies at the process granularity.

[0015] 2. This invention uses time-series synchronization processing based on timestamp alignment rules. For each process data subset, the data source with the highest sampling frequency is selected as the reference time axis, and the data records of other data sources are associated with the corresponding time points on the reference time axis. Furthermore, by extracting the mode value of time deviation as the system deviation compensation amount, a secondary alignment operation is performed. This effectively solves the problem of time-series deviation caused by differences in equipment clocks, data transmission delays, and other factors between different data sources during production execution. It makes the correspondence of multi-source data on a unified time axis accurate to a single sampling period, eliminating the distortion of traceability results caused by time axis misalignment.

[0016] 3. This invention introduces a conflict detection and resolution mechanism based on a set of physical constraints. It uses production process sequence constraints to verify the process node of the same lighting product at the same time, and uses equipment state mutual exclusion constraints to verify the production tasks executed by the same production equipment at the same time. It also corrects or fills in data records that violate the constraints, thus solving the problem of confused process attribution caused by data record conflicts during the fusion process of multi-source databases. This significantly improves the data quality and consistency of multi-source heterogeneous data during the fusion process.

[0017] 4. This invention links standardized process data subsets into a complete production history chain covering the entire process from raw material delivery to finished product warehousing, according to the actual process flow sequence. During the linking process, a judgment mechanism based on shared workstation equipment numbers and time interval thresholds automatically establishes or selects alternative link pointers, ensuring accurate connection between adjacent process data subsets. This achieves complete restoration and efficient retrieval of the entire OLED lighting product production process, providing complete data support for root cause localization when quality anomalies occur. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the overall process of the multi-source database retrieval method for traceability of lighting product production quality in this invention; Figure 2 This is a structural framework diagram of the multi-source database retrieval system for traceability of lighting product production quality according to the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1 like Figure 1 As shown, this invention provides a multi-source database retrieval method for traceability of lighting fixture production quality, including the following steps: S1. In response to the input production batch identifier, locate the corresponding initial data set in multiple preset data sources according to the production batch identifier; wherein, the multiple preset data sources include a first database storing lighting production plan information, a second database storing real-time operating parameters of lighting production equipment, a third database storing batch information of lighting production raw materials, and a fourth database storing lighting production environment parameters. S2. Perform hierarchical classification processing on the initial data set. Based on the production process links associated with the data, the initial data set is decomposed into process data subsets corresponding to multiple production process nodes. Each process data subset contains equipment operation parameter fragments, raw material batch association information, and environmental parameter records under the corresponding production process node, which identify the production batch. S3. For each process data subset, perform time synchronization processing based on timestamp alignment rules, and align the equipment operating parameter fragments, raw material batch association information and environmental parameters from different data sources within the process data subset on a unified time axis to obtain the aligned synchronized process data subset. S4. Perform conflict detection and resolution on the synchronous process data subset: Based on the preset set of physical constraints, traverse the data items in the synchronous process data subset, filter out data items that contradict the preset set of physical constraints, and extract two data items in the synchronous process data subset that are adjacent to the filtered data items at the process node and satisfy the preset set of physical constraints. Use the values ​​of the two data items to perform linear interpolation to fill in the filtered data items, and obtain a standardized process data subset; wherein, the preset set of physical constraints includes production process sequence constraints and equipment state mutual exclusion constraints; S5. Link the standardized process data subsets according to the actual process flow sequence in the lighting production process to form a complete production history chain covering the entire process of lighting products corresponding to the production batch identifier, from raw material input to finished product warehousing, and output the production history chain as the search result.

[0021] Furthermore, the specific implementation method for locating the corresponding initial data set among multiple preset data sources in S1 is as follows: Extract the product model code and production date code segment from the production batch identifier; The product model code is matched with the planned product model field in the first database and the raw material matching model field in the third database, and the production date code is matched with the operation log date field in the second database and the environmental record date field in the fourth database. For multiple successfully matched record entries, extract the data generation timestamps of each record entry, sort the record entries according to the chronological order of the data generation timestamps, and check whether the timestamp interval between adjacent record entries after sorting exceeds the preset sampling interval threshold. If it exceeds the threshold, insert null placeholders between adjacent record entries, and merge all record entries after inserting null placeholders into the initial data set.

[0022] Furthermore, in S2, the specific implementation method for decomposing the initial dataset into process data subsets corresponding to multiple production process nodes based on the production process links associated with the data is as follows: Extract the workstation equipment identification code carried in each data record in the initial dataset; The corresponding process node number is retrieved from the preset process mapping table based on the workstation equipment identification code; the process mapping table stores the correspondence between multiple workstation equipment identification codes and multiple process node numbers; Data records with the same process node number are grouped into the same process data subset; For data records that do not have workstation equipment identification codes, the equipment type code and operation time interval carried by the data record are extracted. Based on the equipment type code, the corresponding candidate process node set is matched from the preset equipment and process mapping table. Then, the operation time interval is compared with the standard operation time window of each process node in the candidate process node set. The candidate process node whose overlap time exceeds the preset overlap threshold is taken as the process node to which the data record belongs, and the data record is assigned to the process data subset corresponding to the process node.

[0023] Furthermore, the timing synchronization processing based on timestamp alignment rules in S3 is specifically as follows: The data source with the highest sampling frequency is selected from the subset of process data as the baseline time axis; For other data sources within the process data subset, identify the original timestamp carried by each data record in the other data sources, find the reference time point on the reference time axis with the smallest time difference from the original timestamp, and associate the data record with the reference time point. Obtain the standard timestamp sequence of each reference time point on the reference time axis, as well as the time deviation value generated when each data record in other data sources is associated with the reference time point. Extract the deviation value with the highest frequency from all time deviation values ​​as the system deviation compensation amount of other data sources. After adding the system deviation compensation amount to the original timestamps of all data records in other data sources, re-execute the reference time point lookup and association operation, and package and mark all the re-associated data records. All the marked data records share the same time axis index value.

[0024] Furthermore, before S2 breaks down the initial dataset into subsets of process data corresponding to multiple production process nodes based on the production process links associated with the data, the following operations are performed: Extract the lamp type code from the production batch identifier, and retrieve the corresponding standard process sequence from the preset process template library according to the lamp type code; the standard process sequence contains multiple standard process nodes arranged according to the actual production sequence of the lamps, and each standard process node carries the standard operation time threshold and standard workstation equipment type set under the lamp type; Obtain the actual workstation equipment type identifier and actual operation time carried by each data record in the initial dataset; Match each actual workstation equipment type identifier with the set of standard workstation equipment types carried by each standard process node, and compare the actual operation time recorded in the data with the standard operation time threshold carried by the standard process node; When the actual workstation equipment type identifier is successfully matched and the actual operation time does not exceed the standard operation time threshold, the data record will be classified into the temporary classification set corresponding to the standard process node. When there are free data records in the initial data set that have not been assigned to any temporary classification set, extract the adjacent process data association markers carried by the free data records, locate the target temporary classification set that has a process adjacency relationship with the free data records from the already classified temporary classification sets based on the adjacent process data association markers, and merge the free data records into the target temporary classification set. All temporary classification sets are output as process data subsets corresponding to multiple production process nodes, according to the arrangement order of the corresponding standard process nodes in the standard process sequence.

[0025] Furthermore, the specific implementation method for forming a production history chain in S5 is as follows: According to the sequence of process flow, the data subsets of each standardized process are arranged into a data link sequence; For two adjacent standardized process data subsets in the data link sequence, extract the set of first station equipment numbers and the set of first process end timestamps carried by all data items in the first standardized process data subset, and extract the set of second station equipment numbers and the set of second process start timestamps carried by all data items in the second standardized process data subset. Perform an intersection operation on the set of equipment numbers for the first workstation and the set of equipment numbers for the second workstation to obtain a subset of shared workstation equipment numbers. Extract the first end timestamp corresponding to the subset of shared workstation equipment numbers from the set of first process end timestamps, and extract the second start timestamp corresponding to the subset of shared workstation equipment numbers from the set of second process start timestamps. Detect whether the time interval between the first end timestamp and the second start timestamp is less than the preset process transition time threshold. When the time interval is less than the threshold for process transition time, a link pointer based on the shared workstation equipment number subset is established between the previous standardized process data subset and the next standardized process data subset. When the time interval is greater than or equal to the threshold of the process transition time, the workstation equipment number pair with the smallest time interval is selected from the set of timestamps at the end of the first process and the set of timestamps at the beginning of the second process. Based on the workstation equipment number pair, an alternative link pointer is established between the data subset of the previous standardized process and the data subset of the next standardized process, and the alternative link pointer is marked as pending manual verification. Iterate through all adjacent relationships, establish all link pointers and alternative link pointers, and output the production history chain with link tags.

[0026] Furthermore, after outputting the production history chain with link tags in S5, the following operations are also performed: Traverse each link pointer and alternative link pointer in the production history chain, extract all workstation equipment number pairs corresponding to the alternative link pointers, and form a sequence of equipment number to be verified. Based on the sequence of equipment numbers to be verified, query the physical location coordinates and equipment type identifiers corresponding to each workstation equipment number from the preset workstation equipment topology diagram; Two workstation equipment numbers that are less than the preset adjacency distance threshold and have the same equipment type identifier are grouped into the same equipment association group. For alternative link pointers within the same device association group, the marking status of the alternative link pointers will be automatically switched from pending manual verification to confirmed status, and the switching record will be retained. For candidate link pointers that are not assigned to any device association group, maintain the status of pending manual verification and output a pending verification prompt message in the production history chain.

[0027] like Figure 2 As shown, the present invention also provides a multi-source database retrieval system for tracing the quality of lighting fixture production, used to execute the multi-source database retrieval method for tracing the quality of lighting fixture production of the present invention, including: The multi-source positioning module is used to respond to the input production batch identifier and locate the corresponding initial data set in multiple preset data sources according to the production batch identifier; wherein, the multiple preset data sources include a first database storing lighting production plan information, a second database storing real-time operating parameters of lighting production equipment, a third database storing batch information of lighting production raw materials, and a fourth database storing lighting production environment parameters. The hierarchical classification module is used to perform hierarchical classification processing on the initial data set. Based on the production process links associated with the data, the initial data set is broken down into process data subsets corresponding to multiple production process nodes. Each process data subset contains equipment operation parameter fragments, raw material batch association information, and environmental parameter records under the corresponding production process node, which identify the production batch. The timing synchronization module is used to perform timing synchronization processing based on timestamp alignment rules for each process data subset. It aligns the equipment operating parameter fragments, raw material batch association information and environmental parameters from different data sources within the process data subset on a unified time axis to obtain the aligned synchronized process data subset. The conflict resolution module is used to perform conflict detection and resolution processing on the synchronous process data subset: Based on a preset set of physical constraints, it traverses the data items in the synchronous process data subset, filters out data items that contradict the preset set of physical constraints, and extracts two data items in the synchronous process data subset that are adjacent to the filtered data items at the process node and satisfy the preset set of physical constraints. The values ​​of the two data items are used to perform linear interpolation to fill in the filtered data items, resulting in a standardized process data subset; wherein, the preset set of physical constraints includes production process sequence constraints and equipment state mutual exclusion constraints; The traceability chain construction module is used to link standardized process data subsets according to the actual process flow sequence in the lighting production process, forming a complete production traceability chain that covers the entire process of lighting products corresponding to the production batch identifier from raw material input to finished product warehousing, and outputs the production traceability chain as search results.

[0028] Furthermore, the multi-source positioning module also integrates a data pre-inspection submodule and a data routing submodule: The data pre-inspection submodule is used to retrieve successfully matched record entries from each preset data source, traverse the data type identifier carried by each record entry, and classify the record entries into structured data items and unstructured data items according to the data type identifier. The structured data items are directly stored in the first buffer of the multi-source positioning module, and the unstructured data items are stored in the second buffer of the multi-source positioning module. The data routing submodule is used to periodically poll the storage status of the first and second caches. When the cumulative storage of structured data items in the first cache reaches a preset batch capacity threshold, all structured data items in the first cache are grouped according to the data generation timestamp of the record entries, generating a structured batch package and pushing it to the hierarchical classification module. When the cumulative storage of unstructured data items in the second cache reaches a preset batch capacity threshold, the key fields of the process steps carried by each unstructured data item are extracted, and all unstructured data items in the second cache are pre-classified according to the key fields of the process steps, generating an unstructured batch package with process step tags and pushing it to the hierarchical classification module.

[0029] Furthermore, the hierarchical classification module also integrates a process node verification submodule and a data correction submodule: The process node verification submodule is used to traverse each process data subset after the hierarchical classification module groups data records with the same process node number into the same process data subset. It checks whether each process data subset contains a complete process start timestamp and process end timestamp. For target data records that are missing process start timestamps or process end timestamps, it extracts the process start timestamps and process end timestamps carried by other data records in the process data subset to which the target data record belongs. It selects the earliest time value from the extracted process start timestamps as the process start timestamp of the target data record, and selects the latest time value from the extracted process end timestamps as the process end timestamp of the target data record. The data correction submodule is used to extract the process node number of the process data subset to which the target data record belongs after the process node verification submodule completes the timestamp completion of the target data record. It then queries the standard operation time range corresponding to the process node number from the preset standard process time mapping table, checks whether the actual operation time between the process start timetamp and the process end timetamp of the target data record falls within the standard operation time range. If it does not fall within the range, the process end timetamp of the target data record is corrected to the sum of the process start timetamp and the standard operation time range, and a correction log record is generated.

[0030] Example 2 Taking a large lighting manufacturing enterprise in the Yangtze River Delta region as an example, the enterprise has multiple product lines such as OLED bulbs, panel lights, and industrial and mining lamps, with a daily output of over 100,000 units. In the process of implementing the multi-source database retrieval method of this invention, the enterprise's multi-source database system consists of an SAP ERP system, a MES system, and a SCADA equipment IoT platform. When the enterprise's quality inspection department finds a batch of panel lights with excessive color temperature deviation during finished product sampling, the quality inspectors input the production batch identifier of the batch of products into the system. The structure of the production batch identifier is P2203A-15, where P2203A represents the product model code and 15 represents the production date code segment, i.e., the 15th day of the current month.

[0031] After receiving the production batch identifier, the multi-source positioning module performs a multi-source positioning operation based on field matching. In the lighting production plan information stored in the first database, the planned product model field matches the product model code P2203A successfully, and outputs the production work order number and planned output of the model for that day. In the third database, the raw material matching model field matches P2203A successfully, and outputs the OLED lamp bead batch number, driver power supply batch number and diffuser plate batch number corresponding to the batch of products. The operation log date field in the second database and the environmental record date field in the fourth database both match the date code segment 15 successfully, and output the equipment operation parameter records and workshop environmental temperature and humidity records associated with the production work order for that day.

[0032] The above-mentioned successfully matched record entries generated a total of 127 data records. After the multi-source positioning module extracted the data of each record entry and generated a timestamp, it sorted these 127 record entries according to their chronological order. Among the adjacent record entries after sorting, the timestamp interval between the end of the equipment operation parameter record and the beginning of the raw material batch information record was 32 seconds, while the preset sampling interval threshold in this scenario was 5 seconds. This timestamp interval clearly exceeded the threshold. The multi-source positioning module inserted null placeholders between the adjacent record entries and merged all record entries into the initial data set.

[0033] After receiving the initial data set, the hierarchical classification module performs a disassembly operation based on the production process links associated with the data. The production process of the company's panel lights consists of six process nodes: SMT assembly, LED chip soldering, driver assembly, optical testing, aging testing, and finished product packaging. The hierarchical classification module extracts the workstation equipment identification code carried by each data record in the initial data set and queries the process node number corresponding to each workstation equipment identification code from the preset process mapping table. There are a total of 118 data records with workstation equipment identification codes, which are classified into the six process data subsets.

[0034] The remaining 9 data records lack workstation equipment identification codes. The hierarchical classification module extracts the equipment type codes and operation time intervals from these 9 data records. Taking one data record as an example, its equipment type code is OPT-03, and its operation time interval is from 10:23 AM to 10:25 AM on the same day. Based on the equipment type code OPT-03, the hierarchical classification module matches the corresponding candidate process node set from the preset equipment-process mapping table. This set includes two candidate process nodes: optical testing and aging testing. The hierarchical classification module further queries the standard operating time window for the optical testing process node, which is 10:20 AM to 10:30 AM, and the aging testing process node... The standard operating time window is from 10:15 to 11:00. After comparing the operation time interval of this data record with the standard operating time windows of the two candidate process nodes, the overlap time with optical testing is 2 minutes and the overlap time with aging testing is 2 minutes. Neither of these overlap times exceeds the preset overlap threshold of 3 minutes. Therefore, this data record is marked as an unclassified free data record. The hierarchical classification module continues to extract the adjacent process data association markers carried by this free data record. These association markers point to the last classified data record of the previous process node driven assembly. Based on these markers, the free data record is located to the temporary classification set corresponding to the optical testing process node.

[0035] After hierarchical classification, the timing synchronization module performs timing synchronization processing on each subset of process data. Taking the optical testing process data subset as an example, this subset includes optical test parameter data sources from the MES system with a sampling frequency of 10Hz, and equipment vibration data sources from the SCADA system with a sampling frequency of 1Hz. The timing synchronization module selects the MES optical test parameter data source with the highest sampling frequency as the reference time axis. This reference time axis contains a total of 120 reference time points with standard timestamp sequences. For each data record in the SCADA equipment vibration data source, the timing synchronization module identifies its original timestamp, finds the reference time point on the reference time axis with the smallest time difference from the original timestamp, and associates this data record with the reference time axis. After the initial association is completed at the reference time point, the time synchronization module obtains the standard timestamp sequence corresponding to all 120 reference time points, as well as the time deviation value generated when each data record in the SCADA equipment vibration data source is associated with the reference time point. The deviation value with the highest frequency is extracted from all time deviation values ​​as the system deviation compensation amount of the data source. In this embodiment, the deviation value with the highest frequency is +0.15s. The time synchronization module adds +0.15s to the original timestamp of all data records in the SCADA equipment vibration data source and then re-executes the reference time point search and association operation. All data records after re-association are packaged and marked so that the MES system data and the SCADA system data share the same time axis index value.

[0036] The conflict resolution module performs conflict detection and resolution processing on the synchronous process data subset based on a preset set of physical constraints. In the LED welding process data subset, it finds that a certain data record indicates that the equipment occupancy time period is from 9:45 AM to 10:05 AM, while another data record of the same welding equipment indicates that the equipment occupancy time period is from 9:50 AM to 9:55 AM. The two time periods overlap and violate the mutual exclusion constraint of equipment status. After marking this data record as the target data record, the conflict resolution module extracts the equipment occupancy time periods from other data records of the same welding equipment as a reference, adjusts the start time of equipment occupancy in the target data record to 10:05 AM, and adjusts the end time of equipment occupancy to the sum of the start time and the original duration of equipment occupancy. The original duration of equipment occupancy is 20 minutes, so the adjusted end time of equipment occupancy is 10:25 AM, thereby eliminating the time overlap. The conflict resolution module outputs the corrected record to the standardized process data subset.

[0037] Example 3 The multi-source database retrieval system for traceability of lighting production quality provided by this invention is deployed on an internal server cluster within the enterprise in terms of hardware architecture. The specific operating logic and working process of each module of the system are as follows.

[0038] The multi-source localization module is the data entry point for the entire system. Internally, it integrates a data pre-detection submodule and a data routing submodule. After retrieving successfully matched record entries from various preset data sources, the data pre-detection submodule iterates through the data type identifiers carried by each record entry. These identifiers are marked through the header information fields of the record entry. Structured data items are represented as tabular data with fixed field lengths, while unstructured data items are represented as log data in JSON or plain text format. Structured data items are stored in the first buffer of the multi-source localization module, and unstructured data items are stored in the second buffer. The data routing submodule polls the storage status of the first and second buffers every 500ms. When the first buffer... When the cumulative storage of structured data items in the storage area reaches a preset batch capacity threshold (in this embodiment, the threshold is 200 record entries), the data routing submodule groups all structured data items in the first cache area according to the data generation timestamp of the record entries, generates a structured batch package, and pushes it to the hierarchical classification module. When the cumulative storage of unstructured data items in the second cache area reaches a preset batch capacity threshold (in this embodiment, the threshold is 100 record entries), the data routing submodule extracts the key fields of the process steps carried by each unstructured data item, performs pre-classification on all unstructured data items in the second cache area according to the key fields of the process steps, generates an unstructured batch package with process step tags, and pushes it to the hierarchical classification module.

[0039] The hierarchical classification module is the core of data decomposition and merging. Internally, it integrates a process node verification submodule and a data correction submodule. After the hierarchical classification module groups data records with the same process node number into the same process data subset, the process node verification submodule iterates through each process data subset, checking whether each subset contains complete process start and end timestamps. Taking the LED bead welding process data subset as an example, this subset contains 45 data records, of which 2 records are missing process end timestamps. The process node verification submodule extracts the process to which these 2 target data records belong. After taking the process start timestamp and process end timestamp carried by the other 43 data records in the sequence data subset, the earliest time value is selected from the extracted process start timestamps as the process start timestamp of the target data record, and the latest time value is selected from the extracted process end timestamps as the process end timestamp of the target data record. In this embodiment, the earliest time value in the extracted process start timestamps is 9:25 AM, and the latest time value in the extracted process end timestamps is 10:08 AM. The process start timestamp of these two target data records is assigned the value of 9:25 AM, and the process end timestamp is assigned the value of 10:08 AM.

[0040] After the process node verification submodule completes the timestamp completion for the target data record, the data correction submodule extracts the process node number of the process data subset as WLD-02. The data correction submodule queries the preset standard process duration mapping table to find the standard operation duration range corresponding to the process node number WLD-02, which is 12min to 18min. The data correction submodule detects that the process start timestamp of the target data record is 9:25 AM and the process end timestamp is 10:08 AM. The actual operation duration between the two is 43min. This actual operation duration does not fall within the standard operation duration range. The data correction submodule corrects the process end timestamp of the target data record to the sum of the process start timestamp and the midpoint of the standard operation duration range. In this embodiment, the midpoint of the standard operation duration range is 15min. Therefore, the corrected process end timestamp is 9:40 AM. The data correction submodule generates a correction log record. The content of the correction log record includes the unique identifier of the target data record, the original process end timestamp, the corrected process end timestamp, the correction reason, and the correction time.

[0041] After receiving the standardized process data subset pushed by the hierarchical classification module, the time-series synchronization module performs a timeline alignment operation on the multi-source data. The time-series synchronization module first scans the number of data records of each data source in the process data subset, identifies the data source with the highest sampling frequency, and sets the timestamp sequence of the data source as the reference timeline. The time-series synchronization module then traverses other data sources, performs a reference time point lookup and association operation on each data record, and after completing all associations, extracts the system deviation compensation amount from the associated records and performs a secondary alignment operation. After completing the secondary alignment, the time-series synchronization module packages and marks all the re-associated data records so that all data records in the process data subset share the same timeline index value. This index value is represented by an auto-incrementing integer sequence, starting from 1 and increasing to the total number of data records contained in the process data subset.

[0042] The conflict resolution module performs physical constraint verification on the synchronized process data subset output by the timing synchronization module. The conflict resolution module pre-stores production process sequence constraints and equipment state mutual exclusion constraints. The production process sequence constraint requires that the same lighting product can only be in one production process node at any given time. This constraint is verified by traversing the timestamp interval of the same product batch identifier in each process node. The equipment state mutual exclusion constraint requires that the same production equipment can only execute one production task at any given time. This constraint is verified by traversing the equipment occupancy time period of the same equipment number in each data record. When a data record that violates the constraints is detected, the conflict resolution module extracts two data items from the process data subset that are adjacent to the filtered data item at the process node and satisfy the constraints. The values ​​of these two data items are used to perform linear interpolation to fill in the filtered data item. After the filling is completed, a standardized process data subset is output.

[0043] The traceability chain construction module is the final step in generating traceability results. It arranges standardized process data subsets into a data chain sequence according to the chronological order of process flow. For two adjacent standardized process data subsets, the module extracts the set of first-station equipment numbers and the set of first-process end timestamps from all data items in the preceding subset, and the set of second-station equipment numbers and the set of second-process start timestamps from all data items in the following subset. The module then performs an intersection operation on the first and second station equipment number sets to obtain a shared station equipment number subset. If the shared station equipment number subset is not empty, it extracts the first end timestamp corresponding to this shared station equipment number subset from the first process end timestamp set and the first end timestamp corresponding to this shared station equipment number subset from the second process start timestamp set. The second start timestamp is used to detect whether the time interval between the first end timestamp and the second start timestamp is less than a preset process transition time threshold. In this embodiment, the threshold is 30s. When the time interval is less than 30s, the history chain construction module establishes a link pointer based on a shared workstation equipment number subset between the previous standardized process data subset and the next standardized process data subset. When the time interval is greater than or equal to 30s, the history chain construction module selects the workstation equipment number pair with the smallest time interval from the first process end timestamp set and the second process start timestamp set. Based on the workstation equipment number pair, it establishes a candidate link pointer between the previous standardized process data subset and the next standardized process data subset, and marks the candidate link pointer as pending manual verification. After the history chain construction module traverses all adjacent relationships and completes the establishment of all link pointers and candidate link pointers, it outputs a production history chain with link markers.

[0044] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

Claims

1. A multi-source database retrieval method for traceability of lighting fixture production quality, characterized in that, Includes the following steps: S1. In response to the input production batch identifier, locate the corresponding initial data set in multiple preset data sources according to the production batch identifier; wherein, the multiple preset data sources include a first database storing lighting production plan information, a second database storing real-time operating parameters of lighting production equipment, a third database storing batch information of lighting production raw materials, and a fourth database storing lighting production environment parameters. S2. Perform hierarchical classification processing on the initial data set. Based on the production process links associated with the data, the initial data set is decomposed into process data subsets corresponding to multiple production process nodes. Each process data subset contains equipment operation parameter fragments, raw material batch association information, and environmental parameter records under the corresponding production process node, which identify the production batch. S3. For each process data subset, perform time synchronization processing based on timestamp alignment rules, and align the equipment operating parameter fragments, raw material batch association information and environmental parameters from different data sources within the process data subset on a unified time axis to obtain the aligned synchronized process data subset. S4. Perform conflict detection and resolution on the synchronous process data subset: Based on the preset set of physical constraints, traverse the data items in the synchronous process data subset, filter out data items that contradict the preset set of physical constraints, and extract two data items in the synchronous process data subset that are adjacent to the filtered data items at the process node and satisfy the preset set of physical constraints. Use the values ​​of the two data items to perform linear interpolation to fill in the filtered data items, and obtain a standardized process data subset; wherein, the preset set of physical constraints includes production process sequence constraints and equipment state mutual exclusion constraints; S5. Link the standardized process data subsets according to the actual process flow sequence in the lighting production process to form a complete production history chain covering the entire process of lighting products corresponding to the production batch identifier, from raw material input to finished product warehousing, and output the production history chain as the search result.

2. The method according to claim 1, characterized in that, The specific implementation method for locating the corresponding initial data set among multiple preset data sources in S1 is as follows: Extract the product model code and production date code segment from the production batch identifier; The product model code is matched with the planned product model field in the first database and the raw material matching model field in the third database, and the production date code is matched with the operation log date field in the second database and the environmental record date field in the fourth database. For multiple successfully matched record entries, extract the data generation timestamps of each record entry, sort the record entries according to the chronological order of the data generation timestamps, and check whether the timestamp interval between adjacent record entries after sorting exceeds the preset sampling interval threshold. If it exceeds the threshold, insert null placeholders between adjacent record entries, and merge all record entries after inserting null placeholders into the initial data set.

3. The method according to claim 1, characterized in that, The specific implementation method of decomposing the initial data set into process data subsets corresponding to multiple production process nodes according to the production process links associated with the data in S2 is as follows: Extract the workstation equipment identification code carried in each data record in the initial dataset; The corresponding process node number is retrieved from the preset process mapping table based on the workstation equipment identification code; the process mapping table stores the correspondence between multiple workstation equipment identification codes and multiple process node numbers; Data records with the same process node number are grouped into the same process data subset; For data records that do not have workstation equipment identification codes, the equipment type code and operation time interval carried by the data record are extracted. Based on the equipment type code, the corresponding candidate process node set is matched from the preset equipment and process mapping table. Then, the operation time interval is compared with the standard operation time window of each process node in the candidate process node set. The candidate process node whose overlap time exceeds the preset overlap threshold is taken as the process node to which the data record belongs, and the data record is assigned to the process data subset corresponding to the process node.

4. The method according to claim 1, characterized in that, The timing synchronization process based on timestamp alignment rules in S3 specifically includes: The data source with the highest sampling frequency is selected from the subset of process data as the baseline time axis; For other data sources within the process data subset, identify the original timestamp carried by each data record in the other data sources, find the reference time point on the reference time axis with the smallest time difference from the original timestamp, and associate the data record with the reference time point. Obtain the standard timestamp sequence of each reference time point on the reference time axis, as well as the time deviation value generated when each data record in other data sources is associated with the reference time point. Extract the deviation value with the highest frequency from all time deviation values ​​as the system deviation compensation amount of other data sources. After adding the system deviation compensation amount to the original timestamps of all data records in other data sources, re-execute the reference time point lookup and association operation, and package and mark all the re-associated data records. All the marked data records share the same time axis index value.

5. The method according to claim 1, characterized in that, Before decomposing the initial data set into process data subsets corresponding to multiple production process nodes based on the production process links associated with the data in step S2, the following operations are also performed: Extract the lamp type code from the production batch identifier, and retrieve the corresponding standard process sequence from the preset process template library according to the lamp type code; the standard process sequence contains multiple standard process nodes arranged according to the actual production sequence of the lamps, and each standard process node carries the standard operation time threshold and standard workstation equipment type set under the lamp type; Obtain the actual workstation equipment type identifier and actual operation time carried by each data record in the initial dataset; Match each actual workstation equipment type identifier with the set of standard workstation equipment types carried by each standard process node, and compare the actual operation time recorded in the data with the standard operation time threshold carried by the standard process node; When the actual workstation equipment type identifier is successfully matched and the actual operation time does not exceed the standard operation time threshold, the data record will be classified into the temporary classification set corresponding to the standard process node. When there are free data records in the initial data set that have not been assigned to any temporary classification set, extract the adjacent process data association markers carried by the free data records, locate the target temporary classification set that has a process adjacency relationship with the free data records from the already classified temporary classification sets based on the adjacent process data association markers, and merge the free data records into the target temporary classification set. All temporary classification sets are output as process data subsets corresponding to multiple production process nodes, according to the arrangement order of the corresponding standard process nodes in the standard process sequence.

6. The method according to claim 1, characterized in that, The specific implementation method for forming the production history chain in S5 is as follows: According to the sequence of process flow, the data subsets of each standardized process are arranged into a data link sequence; For two adjacent standardized process data subsets in the data link sequence, extract the set of first station equipment numbers and the set of first process end timestamps carried by all data items in the first standardized process data subset, and extract the set of second station equipment numbers and the set of second process start timestamps carried by all data items in the second standardized process data subset. Perform an intersection operation on the set of equipment numbers for the first workstation and the set of equipment numbers for the second workstation to obtain a subset of shared workstation equipment numbers. Extract the first end timestamp corresponding to the subset of shared workstation equipment numbers from the set of first process end timestamps, and extract the second start timestamp corresponding to the subset of shared workstation equipment numbers from the set of second process start timestamps. Detect whether the time interval between the first end timestamp and the second start timestamp is less than the preset process transition time threshold. When the time interval is less than the threshold for process transition time, a link pointer based on the shared workstation equipment number subset is established between the previous standardized process data subset and the next standardized process data subset. When the time interval is greater than or equal to the threshold of the process transition time, the workstation equipment number pair with the smallest time interval is selected from the set of timestamps at the end of the first process and the set of timestamps at the beginning of the second process. Based on the workstation equipment number pair, an alternative link pointer is established between the data subset of the previous standardized process and the data subset of the next standardized process, and the alternative link pointer is marked as pending manual verification. Iterate through all adjacent relationships, establish all link pointers and alternative link pointers, and output the production history chain with link tags.

7. The method according to claim 6, characterized in that, After outputting the production history chain with link tags in S5, the following operations are also performed: Traverse each link pointer and alternative link pointer in the production history chain, extract all workstation equipment number pairs corresponding to the alternative link pointers, and form a sequence of equipment number to be verified. Based on the sequence of equipment numbers to be verified, query the physical location coordinates and equipment type identifiers corresponding to each workstation equipment number from the preset workstation equipment topology diagram; Two workstation equipment numbers that are less than the preset adjacency distance threshold and have the same equipment type identifier are grouped into the same equipment association group. For alternative link pointers within the same device association group, the marking status of the alternative link pointers will be automatically switched from pending manual verification to confirmed status, and the switching record will be retained. For candidate link pointers that are not assigned to any device association group, maintain the status of pending manual verification and output a pending verification prompt message in the production history chain.

8. A multi-source database retrieval system for traceability of lighting fixture production quality, used to execute the method according to any one of claims 1-7, characterized in that, include: The multi-source positioning module is used to respond to the input production batch identifier and locate the corresponding initial data set in multiple preset data sources according to the production batch identifier; wherein, the multiple preset data sources include a first database storing lighting production plan information, a second database storing real-time operating parameters of lighting production equipment, a third database storing batch information of lighting production raw materials, and a fourth database storing lighting production environment parameters. The hierarchical classification module is used to perform hierarchical classification processing on the initial data set. Based on the production process links associated with the data, the initial data set is broken down into process data subsets corresponding to multiple production process nodes. Each process data subset contains equipment operation parameter fragments, raw material batch association information, and environmental parameter records under the corresponding production process node, which identify the production batch. The timing synchronization module is used to perform timing synchronization processing based on timestamp alignment rules for each process data subset. It aligns the equipment operating parameter fragments, raw material batch association information and environmental parameters from different data sources within the process data subset on a unified time axis to obtain the aligned synchronized process data subset. The conflict resolution module is used to perform conflict detection and resolution processing on the synchronous process data subset: Based on a preset set of physical constraints, it traverses the data items in the synchronous process data subset, filters out data items that contradict the preset set of physical constraints, and extracts two data items in the synchronous process data subset that are adjacent to the filtered data items at the process node and satisfy the preset set of physical constraints. The values ​​of the two data items are used to perform linear interpolation to fill in the filtered data items, resulting in a standardized process data subset; wherein, the preset set of physical constraints includes production process sequence constraints and equipment state mutual exclusion constraints; The traceability chain construction module is used to link standardized process data subsets according to the actual process flow sequence in the lighting production process, forming a complete production traceability chain that covers the entire process of lighting products corresponding to the production batch identifier from raw material input to finished product warehousing, and outputs the production traceability chain as search results.

9. The system according to claim 8, characterized in that, The multi-source positioning module also integrates a data pre-inspection submodule and a data routing submodule: The data pre-inspection submodule is used to retrieve successfully matched record entries from each preset data source, traverse the data type identifier carried by each record entry, and classify the record entries into structured data items and unstructured data items according to the data type identifier. The structured data items are directly stored in the first buffer of the multi-source positioning module, and the unstructured data items are stored in the second buffer of the multi-source positioning module. The data routing submodule is used to periodically poll the storage status of the first cache and the second cache. When it is detected that the cumulative storage amount of structured data items in the first cache reaches the preset batch capacity threshold, all structured data items in the first cache are grouped according to the data generation timestamp of the record item, a structured batch package is generated and pushed to the hierarchical classification module. When the cumulative storage of unstructured data items in the second buffer reaches the preset batch capacity threshold, the key fields of the process steps carried by each unstructured data item are extracted. All unstructured data items in the second buffer are pre-classified according to the key fields of the process steps, and an unstructured batch package with process step tags is generated and pushed to the hierarchical classification module.

10. The system according to claim 8, characterized in that, The hierarchical classification module also integrates a process node verification submodule and a data correction submodule. The process node verification submodule is used to traverse each process data subset after the hierarchical classification module groups data records with the same process node number into the same process data subset. It checks whether each process data subset contains a complete process start timestamp and process end timestamp. For target data records that are missing process start timestamps or process end timestamps, it extracts the process start timestamps and process end timestamps carried by other data records in the process data subset to which the target data record belongs. It selects the earliest time value from the extracted process start timestamps as the process start timestamp of the target data record, and selects the latest time value from the extracted process end timestamps as the process end timestamp of the target data record. The data correction submodule is used to extract the process node number of the process data subset to which the target data record belongs after the process node verification submodule completes the timestamp completion of the target data record. It then queries the standard operation time range corresponding to the process node number from the preset standard process time mapping table, checks whether the actual operation time between the process start timetamp and the process end timetamp of the target data record falls within the standard operation time range. If it does not fall within the range, the process end timetamp of the target data record is corrected to the sum of the process start timetamp and the standard operation time range, and a correction log record is generated.

Images