What the data for this category looks like
Due diligence data sources include operating ledgers from internal operating entities, publicly available industrial and commercial registration information, publicly disclosed industry regulatory documents, and internal audit working papers. Update schedules follow three patterns: quarterly updated operating entity data, monthly updated related party transaction records, and irregularly updated regulatory disclosure documents. Document structure includes three types: group-level overall due diligence summary documents, independent due diligence dossiers for each subsidiary, and structured related party relationship datasets. Fields covered include entity name, establishment date, registered capital, related party transaction amount, and related party shareholding records. Registered capital uses ten thousand yuan as the unit. Related party transaction amount uses yuan as the unit.
What constraints these characteristics impose on workflow orchestration
Multi-source heterogeneous data sources require workflows to be configured with multiple data source access nodes. These nodes adapt to different formats of operating ledgers, industrial and commercial public documents, and audit working papers. Data sources with different update schedules must be bound to differentiated trigger rules. This prevents redundant synchronization or delayed updates. Single due diligence documents have large volume, including multiple subsidiary dossiers and related party relationship data. Workflows must be configured with segmented parsing and merged output nodes to prevent execution timeouts. Structured related party relationship data requires preset field mapping rules. These rules unify entity identification fields from different sources, avoiding due diligence content deviations caused by matching errors.
Configuration Settings
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
workflow_parallel_limit | 8 parallel tasks | A large number of operating entities require a reasonable concurrency limit to prevent triggering external system rate limits |
parse_file_timeout | 900 seconds | Single due diligence documents include multiple subsidiary dossiers. Sufficient time is required to complete full parsing |
global_variable_scope | Request-level isolation | Global variables for different due diligence tasks do not interfere with each other. This prevents report errors caused by data cross-use |
trigger_type | Combined scheduled trigger + event trigger | Adapts to differentiated update schedules for quarterly reports, monthly related party transactions, and irregular regulatory documents |
field_mapping_rule | Unique matching by entity name | Unifies entity identification fields with different naming conventions across multiple data sources. This ensures accurate matching of related party relationships |
workflow_output_filter | Retain core due diligence fields | Focuses on core content such as related party transactions and shareholding status. Filters redundant information to simplify report structure |
The parameter values provided on this page are common starting points for configuration. Actual values are affected by material format, data volume, and business rules. Specific issues require specific analysis. It is recommended to test on your own samples before finalizing settings.
Three Common Mistakes
- Phenomenon: Global variable values are overwritten by subsequent tasks after workflow runs. Results of different due diligence tasks are mixed together. Cause: Global variable isolation rules are not configured correctly. Globally shared variable storage methods are used instead.
- Phenomenon: Judgment node trigger logic operates abnormally. Results that meet "equal to" or "starts with" conditions do not enter the IF branch. Instead, they enter the ELSE branch. Cause: The binding logic between judgment rules and branches is not understood correctly. The corresponding relationship between rules and branches is reversed.
- Phenomenon: Workflows cannot obtain parameters passed by the system. Core fields are empty or trigger errors occur. Cause: System parameter mapping is not configured in workflow nodes. System parameters are not correctly written to workflow input items or global variables.
How to Confirm Proper Configuration
- Run a single test due diligence document. Verify that global variable values only take effect within the current task and are not interfered with by other parallel tasks.
- Configure different judgment rules, manually input test text, and verify that judgment branch triggers match expected logic.
- Import a test request containing system parameters. Check if the workflow can correctly obtain and use these parameters to generate corresponding content.
- Run a test document containing multiple subsidiary data. Verify that parallel node task execution operates normally, and that output results fully cover all entity data.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-14.